Introduction

An antisense transcript to the miR34a HG was detected using the UCSC genome browser.

Methods

The USCS genome browser was utilized for the bioinformatic evaluation of antisense transcription at the miR34a locus.

Results

url <- fileMap(type = "png")["Figure 1a"][[1]]
knitr::include_graphics(file.path(projectUrl, url))

Conclusion

Refseq annotation supports the presence of a transcript orininating in the antisense orientation from the miR34a gene.

Introduction

We wished to examine expression of miR34a asRNA in different cancer tissues and, therefore, utilized a panel of common cancer cell lines to detect miR34a asRNA expression, as well as, miR34a HG. We included both p53 wild type, p53 mutated, and p53 null cell lines in the cell line panel due to the fact that miR34a is a known downstream target of p53.

Methods

Cell culture and PCR

Cells were cultured with the appropriate growth medium and seeded at a desity between 50-75% confluency. The following day, RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. PCR was run for 35 or 38 or 16 cycles for miR34a asRNA, miR34a HG, and beta-actin, respectivley, with 58˚C annealing temperature. PCR products were analyzed on a 2% agarose gel.

Primers

primers <- data.frame(
    name=c(
        "miR34a asRNA F1",
        "miR34a asRNA R1",
        "miR34aHG_F",
        "miR34aHG_R",
        "B-actin_F",
        "B-actin_R"
    ),
    sequence=c(
        "AGC GGC ATC TCC TCC ACC TGA AA",
        "TTG CCT CGT GAG TCC AAG GAG AAT",
        "TCT GCT CCA GTG GCT GAT GAG AAA",
        "GTT CAC TGG CCT CAA AGT TGG CAT",
        "AGG TCA TCA CCA TTG GCA ATG AG",
        "CTT TGC GGA TGT CCA CGT CA"
    )
)
primers

Results

url <- fileMap(type = "png")["Figure 1b"][[1]]
knitr::include_graphics(file.path(projectUrl, url))

Conclusions

We were able to detect co-expression of both miR34a HG and miR34a AS in a variety of cell lines, including HCT116 human colon cancer cells and 293T human embryonic kidney cells.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] gtable_0.2.0                  ggthemes_3.4.0               
##  [3] printr_0.1                    forcats_0.2.0                
##  [5] stringr_1.2.0                 dplyr_0.7.4                  
##  [7] purrr_0.2.4                   readr_1.1.1                  
##  [9] tidyr_0.8.0                   tibble_1.4.2                 
## [11] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [13] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         Matrix_1.2-12          rmarkdown_1.9         
## [25] liftr_0.8              splines_3.4.3          selectr_0.3-1         
## [28] foreign_0.8-69         ansistrings_1.0.0.9000 munsell_0.4.3         
## [31] rgeolocate_1.0.1       broom_0.4.3            compiler_3.4.3        
## [34] modelr_0.1.1           pkgconfig_2.0.1        mnormt_1.5-5          
## [37] htmltools_0.3.6        XML_3.98-1.9           crayon_1.3.4          
## [40] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [43] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [46] stringi_1.1.6          reshape2_1.4.3         bindrcpp_0.2          
## [49] xml2_1.2.0             rematch2_2.0.1         tools_3.4.3           
## [52] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [55] survival_2.41-3        yaml_2.1.16            colorspace_1.3-2      
## [58] rvest_0.3.2            bindr_0.1              haven_1.1.1

Introduction

We interogated RNA sequencing data from TCGA to determine in miR34a and miR34a asRNA expression was correlated in both TP53 wild type and mutated samples.

Methods

RNA-Seq data and copy number data were downloaded from TCGA and processed as described previously (Ashouri et al. 2016). Briefly, RNA-Seq data were aligned to the human hg19 assembly and quantified using GENCODE (v19) annotated HTSeq-counts and FPKM normalizations. Expression data from miR34a and miR34 asRNA (identified as RP3-510D11.2) were used for further analysis. Copy number amplitudes for GENCODE genes were determined from segmented copy-number data. Samples that were diploid for miR34 asRNA were identified as those samples that had copy number amplitudes between -0.1 and 0.1.

Somatic mutation data were downloaded from the Genomics Data Commons data portal (GDC) as mutation annotation format (maf) files, called using Mutect2 on 30/10/2017 (v7) (Grossman et al. 2016).

Results

data <- getData("Figure 1c")

#normalize expression values
data <- mutate(
    data,
    RP3 = log2(RP3 / max(RP3)),
    miR34a = log2(miR34a / max(miR34a))  
  )

#add BRCA subtypes
data <- data %>%
  mutate(PAM50 = if_else(is.na(PAM50), "", PAM50)) %>%
  mutate(cancer_PAM50 = gsub("BRCA", "BRCA ", paste(cancer, PAM50, sep = "")))

#sort 
data <- arrange(data, cancer_PAM50, TP53, RP3)

#remove infinite data
data <- filter(data, !is.infinite(RP3) & !is.infinite(miR34a))

#diploid only
data <- filter(data, abs(RP3_cna) < 0.1)

#plot
plotTCGAcorrelation(data, type = "diploid only")

##save to pdf
# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 1c"][[1]])
# pdf(path, width = 7.7, height = 4)
# plotTCGAcorrelation(data, type = "diploid only")
# invisible(dev.off())

See Supplementary Figure 1a for R-squared and p-values corersponding with the analysis.

Conclusions

miR34a asRNA and miR34a tend to be highly correlated in the cancer types examined in both TP53 wild type and mutated samples. In addition, the results indicate that both miR34a asRNA and miR34a expression levels are decreased in the presence of TP53 mutations.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         Matrix_1.2-12          rmarkdown_1.9         
## [25] liftr_0.8              splines_3.4.3          selectr_0.3-1         
## [28] foreign_0.8-69         ansistrings_1.0.0.9000 munsell_0.4.3         
## [31] rgeolocate_1.0.1       broom_0.4.3            compiler_3.4.3        
## [34] modelr_0.1.1           pkgconfig_2.0.1        mnormt_1.5-5          
## [37] htmltools_0.3.6        XML_3.98-1.9           crayon_1.3.4          
## [40] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [43] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [46] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [49] rematch2_2.0.1         tools_3.4.3            glue_1.2.0            
## [52] hms_0.4.1              parallel_3.4.3         survival_2.41-3       
## [55] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [58] bindr_0.1              haven_1.1.1

Introduction

In order to define the 3’ transcription termination site for the miR34a asRNA, 3’-RACE was performed.

Methods

Cell culture and 3-prime RACE

All cell lines were cultured at 5% CO2 and 37° C with U2OS cells grown in McCoy’s 5a (Life Technologies). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. U2OS cell RNA was polyA-tailed using yeast polyA polymerase after which cDNA was synthesized using oligo(dT) primers. Nested-PCR was performed first using a forward primer in miR34a asRNA exon 1 (miR34a_asRNA_F10) and a tailed oligo(dT) primer (polyT_T7FAM) followed by a second PCR using an alternate miR34a asRNA exon 1 primer (miR34a_asRNA_F1) and a reverse primer binding to the tail of the previously used oligo(dT) primer (FAMprimer). PCR products were gel purified and cloned the Strata Clone Kit (Agilent Technologies), and sequenced.

Primers

primers <- data.frame(
  name = c("miR34a_asRNA_F10", "polyT_T7FAM", "miR34a_asRNA_F1", "FAMprimer"),
  sequence = c(
    "ACG CGT CTC TCC AGC CCG GGA T",
    "CAG TGA ATT GTA ATA CGA CTC ACT ATA GGG ACA TCC GTA GCT CGT CCA GGA CCC TTT TTT TTT TTT TTT TTT VN",
    "AGC GGC ATC TCC TCC ACC TGA AA",
    "CCG TAG CTC GTC CAG GAC CC"
  )
)
primers

data <- tibble(
  gene = readr::parse_factor(
    c(rep("miR34a asRNA", 5), rep("lnc34a", 3), "miR34a asRNA F1", "miR34a asRNA F10"),
    levels = c("miR34a asRNA", "lnc34a", "miR34a asRNA F1", "miR34a asRNA F10")
  ),
  feature = c(
    "exon", "intron", "exon", "intron", "exon",
    "exon", "intron", "exon",
    "primer", "primer"
  ),
  start = c(
    9181991, 9182316, 9183633, 9183807, 9191393,
    9181737, 9182295, 9196912,
    9182279, 9182223
    ),
  stop = c(
    9182316, 9183633, 9183807, 9191393, 9191686,
    9182295, 9196912, 9197043,
    9182301, 9182244
  )
)

p <- ggplot(data = NULL) +
  gggenes::geom_gene_arrow(
    data = filter(data, feature %in% c("exon") & gene == "miR34a asRNA"),
    aes(xmin = start, xmax = stop, y = gene, fill = gene),
    arrowhead_width = unit(0, "mm"),
    arrowhead_height = unit(0, "mm"),
    colour = "gray"
  )  +
  geom_segment(
    data = filter(data, feature == "intron" & gene == "miR34a asRNA"),
    aes(x = start, xend = stop, y = gene, yend = gene),
    linetype = "dotted",
    colour = "gray",
    size = 1
  ) +
  geom_segment(
    data = filter(data, feature == "primer"),
    aes(x = start, xend = stop, y = 1.1, yend = 1.1, colour = gene),
    size = 5 
  ) +
  labs(x = "Chromosome 1") +
  guides(fill = FALSE, colour = guide_legend(title = "Primer"))

plotRmarkdown(p) +
  theme_void() +
  scale_fill_brewer()

Results

url <- fileMap(type = "png")["Figure 1d"][[1]]
knitr::include_graphics(file.path(projectUrl, url))

Conclusions

Sequencing of the resulting cloned cDNA indicated the transcripts 3’ transcription termination site to be 125bp upstream of the RP3-510D11.2 transcript’s annotated termination site.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] survival_2.41-3        yaml_2.1.16            colorspace_1.3-2      
## [61] rvest_0.3.2            bindr_0.1              haven_1.1.1

Introduction

We desired to determine the 5’-start site for the miR34a asRNA.

Methods

Cell culture and PCR

To accomplish this we performed the primer walk assay. All cell lines were cultured at 5% CO2 and 37° C with HEK293T cells cultured in DMEM high glucose (Hyclone). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. PCR was performed for 35 cycles at 62°C annealing temperature and the resulting products were analyzed on a 2% agarose gel.

Primers

primers <- data.frame(
    name=c(
        "miR34aAS_F10",
        "miR34aAS_F11",
        "miR34aAS_F12",
        "miR34aAS_F13",
        "miR34aAS_F14",
        "miR34aAS_F15",
        "miR34aAS_R1"
    ),
    sequence=c(
        "ACG CGT CTC TCC AGC CCG GGA T",
        "ATC TGC GTG GTC ACC GAG AAG CA",
        "CGC ACG GAC TGA GAA ACA CAA G",
        "ACG GAG GCT ACA CAA TTG AAC AGG",
        "AGG GAA GAA AGA ACT AGC CGA GCA",
        "CAT TTG CTG CAA TAT CAC CGT GGC",
        "TTG CCT CGT GAG TCC AAG GAG AAT"
    )
)
primers

Results

Primer placement at the locus can be viewed in Supplementary Figure 2a.

url <- fileMap(type = "png")["Figure 1e"][[1]]
knitr::include_graphics(file.path(projectUrl, url))

Conclusion

It appears as though the miR34a asRNA transcript start site is somewhat upstream of the annotated start site. Although there is some product present for the F12 primer, the majority of transcription seems to be originating from the F11 primer. This indicates that the start site is between 90-224bp upstream of the annotated start site.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] survival_2.41-3        yaml_2.1.16            colorspace_1.3-2      
## [61] rvest_0.3.2            bindr_0.1              haven_1.1.1

Introduction

We utilized a bioinformatic approach to evaluate the coding potential of the miR34a asRNA transcript. The Coding-potential assessment tool uses a linear regression model to evaluate coding-potential by examining ORF length, ORF coverage, Fickett score and hexamer score. We further confirmed these results using the Coding-potential Calculator which utilizes a support based machine-based classifier and accesses an alternate set of discriminatory features.

Methods

Protein-coding capacity was evaluated using the Coding-potential Assessment Tool and Coding-potential Calculator with default settings. Transcript sequences for use with Coding-potential Assessment Tool were downloaded from the UCSC genome browser using the following IDs: HOTAIR (ENST00000455246.1), XIST (ENST00000429829.1), β-actin (ENST00000331789.5), Tubulin (ENST00000427480.1), and MYC (ENST00000377970). Transcript sequences for use with Coding-potential Calculator were downloaded from the UCSC genome browser using the following IDs: HOTAIR (uc031qho.1), β-actin (uc003soq.4).

Results

Coding Potential Assesment Tool

cpat <- getData("Figure 1f")
cpat

*Results for Coding-potential Calculator are located in Supplementary Figure 2e.

Conclusions

Results indicated that miR34a asRNA has a similar lack of coding capacity to the known non-coding transcripts HOTAIR and XIST and differs greatly when examining these parameters to the known coding transcripts beta-actin, tubulin, and MYC. However, to fully evaluate coding potential methods such as mass spectrometry or ribosome profiling must be used.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] survival_2.41-3        yaml_2.1.16            colorspace_1.3-2      
## [61] rvest_0.3.2            bindr_0.1              haven_1.1.1

Introduction

miR34a is a known downstream target of p53 and has been previously shown to exhibit increased expression with cellular apoptotic signaling. We hypothesized that miR34a asRNA may be regulated in a similar fashion whereby transcription is stimulated by activation of p53. To test this we treated HCT116 and HEK293t cells with 200 ng/ml of the DNA damaging agent doxorubicin for 24 hours and monitored miR34a asRNA expression.

Methods

Cell culture and QPCR

All cell lines were cultured at 5% CO2 and 37° C with HEK293T cells cultured in DMEM high glucose (Hyclone) and HCT116 cells in McCoy’s 5a (Life Technologies). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. Cells were plated at 300,000 cells per well in a 6-well plate and cultured overnight. The following day cells were treated with 0, 100, 200, or 500 ng/ml doxorubicin for 24hrs. RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. QPCR was carried out using KAPA 2G SYBRGreen (Kapa Biosystems) using the Applied Biosystems 7900HT machine with the cycling conditions: 95 °C for 3 min, 95 °C for 3 s, 60 °C for 30 s. qPCR primers are shown below.

Analysis

Two experimental (technical) replicates were included in each QPCR run and delta ct was calculated for each sample using the mean of the gene of interest’s technical replicates and the house keeping gene’s technical replicates. delta-delta ct was calculated for each sample by subtracting the median dct value for the corresponding untreated samples. Three independant experiments (biological replicates) were performed in total. The Student’s t-test was used to compare the treated vs untreated group’s delta-delta ct values for both genes.

Primers

primers <- data.frame(
    name=c(
        "ß-actin Fwd",
        "ß-actin Rev",
        "miR34a HG_F",
        "miR34a HG_R",
        "miR34a_asF1",
        "miR34a_asR1"
        
    ),
    sequence=c(
        "AGGTCATCACCATTGGCAATGAG",
        "CTTTCGGGATGTCCACGTCA",
        "TCTGCTCCAGTGGCTGATGAGAAA",
        "GTTCACTGGCCTCAAAGTTGGCAT",
        "AGCGGCATCTCCTCCACCTGAAA",
        "TTGCCTCGTGAGTCCAAGGAGAAT"
        
    )
)
primers

Results

data <- getData("Figure 2a")

#calculate dct
groups <- c("Treatment", "Biological Replicate", "Cell line", "gene")
gois <- c("lncTAM34a", "miR34a HG")

data <- technicalMeans(data, groups) %>%
  dct(., gois, Actin, groups)

#calculate ddct
groups[groups == "gene"] <- "GOI"
logical <- tibble(
  `Biological Replicate` = pull(data, `Biological Replicate`) == 1,
  Treatment = pull(data, Treatment) == "untreated"
)

data <- data %>%
  ddct(., logical, groups) %>%
  folds(.)

#calculate stats
stats <- calcStats(data, Treatment, "untreated", groups) %>%
  pFormat(.)

#prepare for plotting

#vertical lines and labels
vl <- stats %>%
  group_by(`Cell line`, GOI) %>%
  summarize(
    max = max(CI95h),
    pValue = as.numeric(min(pValue))
  ) %>%
  ungroup() %>%
  mutate(max = max + c(0, 0.4, 0, 0)) %>%
  pFormat(.) %>%
  mutate(
    x = case_when(
      GOI == "lncTAM34a" ~ 1 - 0.25,
      GOI ==    "miR34a HG" ~ 1 + 0.25
    ),
    xend = case_when(
      GOI == "lncTAM34a" ~ 2 - 0.25,
      GOI ==    "miR34a HG" ~ 2 + 0.25
    ),
    pFormat = pFormat,
    labelX = 1.5
  )

data <- data %>%
  mutate(Treatment = case_when(
    Treatment == "untreated" ~ 0,
    Treatment == "doxorubicin" ~ 200,
    TRUE ~ NaN
  )) %>%
  mutate(Treatment = parse_factor(Treatment, levels = c("0", "200")))

stats <- stats %>%
  mutate(Treatment = case_when(
    Treatment == "untreated" ~ 0,
    Treatment == "doxorubicin" ~ 200,
    TRUE ~ NaN
  )) %>%
  mutate(Treatment = parse_factor(Treatment, levels = c("0", "200")))

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = Treatment,
      y = log2fold,
      fill = GOI
    ),
    binwidth = 1/10,
    colour = NA,
    #alpha = 0.5,
    position = position_dodge(width = 0.9), 
    binaxis='y', 
    stackdir='center',
    dotsize = 3,
    show.legend = TRUE
  ) +
  facet_grid(. ~ `Cell line`) +
  scale_y_continuous(
    "\u0394\u0394Ct",
    breaks=c(-1,0,1,2,3,4,5)
  ) +
  labs(
    x = "Doxorubicin (ng/ml)",
    title = "miR34a asRNA response to p53 induction"
  ) +
  guides(
    fill = guide_legend(
      title = "Gene", 
      override.aes = list(size = 6)
  ))

markdown <- p +
  geom_point(
    data = stats,
    aes(
      x = Treatment,
      y = mean,
      group = GOI
    ),
    size = 6,
    position = position_dodge(width = 0.9),
    show.legend = FALSE,
    shape = 95,
    colour = "gray20"
  ) +
  geom_linerange(
    data = stats,
    aes(
      x = Treatment,
      ymin = CI95l,
      ymax = CI95h,
      group = GOI
    ),
    colour = "gray20",
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_label(
    data = vl,
    aes(
      x = labelX,
      y = max + 0.25,
      label = pFormat,
      group = GOI
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 5,
    family = "Arial Unicode MS",
    position = position_dodge(width = 0.9)
  ) +
  geom_segment(
    data = vl,
    aes(
      x = x,
      y = max + 0.1,
      xend = xend,
      yend = max + 0.1,
      group = GOI
    ),
    colour="grey43",
    show.legend = FALSE
  ) +
  guides(
    fill = guide_legend(
      title = "Gene", 
      override.aes = list(size = 6)
  ))

plotRmarkdown(markdown)

pdf <- p + 
  geom_point(
    data = stats,
    aes(
      x = Treatment,
      y = mean,
      group = GOI
    ),
    position = position_dodge(width = 0.9),
    show.legend = FALSE,
    shape = 95,
    size = 4,
    colour = "grey30"
  ) +
  geom_linerange(
    data = stats,
    aes(
      x = Treatment,
      ymin = CI95l,
      ymax = CI95h,
      group = GOI
    ),
    colour = "grey30",
    size = 0.3,
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_label(
    data = vl,
    aes(
      x = labelX,
      y = max + 0.5,
      label = pFormat,
      group = GOI
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    colour = "grey30",
    size = 2.5,
    family = "Arial Unicode MS",
    position = position_dodge(width = 0.9)
  ) +
  geom_segment(
    data = vl,
    aes(
      x = x,
      y = max + 0.1,
      xend = xend,
      yend = max + 0.1,
      group = GOI
    ),
    colour="grey30",
    size = 0.3,
    show.legend = FALSE
  ) +
  guides(
    fill = guide_legend(
      title = "Gene", 
      override.aes = list(size = 4),
      ncol = 1
  ))

figure <- plotPDF(pdf)

path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 2a"][[1]])
ggsave(
   plot = figure,
   filename = path,
   device = cairo_pdf,
   height = 60,
   width = 53,
   units = "mm"
)

Conclusions

We observed increases in miR34a HG and asRNA expression upon doxorubicin treatment indicating that these two transcripts are co-regulated.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] tidyselect_0.2.3       ggfittext_0.5.0        XML_3.98-1.9          
## [43] crayon_1.3.4           withr_2.1.1.9000       nlme_3.1-131          
## [46] jsonlite_1.5           magrittr_1.5           scales_0.5.0.9000     
## [49] cli_1.0.0.9001         stringi_1.1.6          reshape2_1.4.3        
## [52] xml2_1.2.0             RColorBrewer_1.1-2     rematch2_2.0.1        
## [55] tools_3.4.3            glue_1.2.0             hms_0.4.1             
## [58] parallel_3.4.3         survival_2.41-3        yaml_2.1.16           
## [61] colorspace_1.3-2       rvest_0.3.2            bindr_0.1             
## [64] haven_1.1.1

Introduction

miR34a is a known downstream target of p53 and has been previously shown to exhibit increased expression upon cellular apoptotic signaling. We hypothesized that miR34a asRNA may be regulated in a similar fashion whereby transcription is stimulated by activation of p53. To test this we treated p53 wild type and p53 null HCT116 cells with increasing concentrations of the DNA damaging agent doxorubicin and monitored miR34a asRNA expression.

Methods

Cell culture and QPCR

HCT116 cells were cultured in DMEM high modified (Hyclone, GE healthcare) supplemented with supplemented with 2mM L-glutamine, 50ug/ml Penicillin-Streptomycin and 10% Fetal Calf Serum. 25 x 10^5 HCT116 cells were plated in 6 well plates. 24 hours later media was exchanged and doxorubicin was added to a final concentration of 100, 200 or 500 ng/ml. Cells were harvested for RNA extraction 24 hours later using trypsin. RNA was extracted using Nucleospin RNA kit (Machery-Nagel Ref. 740955) according to manufacturer‟s protocol and DNase treated using Ambion Turbo DNA-free according to manufacturer‟s protocol (Life Technologies Ref. AM1907). cDNA was synthesized using ~500 ng RNA with M-MLV (Life Technologies Ref 28025013) and a mixture of oligo(dT) with nanomers in accordance with the manufacturer’s protocol. qPCR quantification was carried out using the PowerUp SYBR Green Master Mix (Thermo Fisher Scientific, Ref. A25777) on the CFX96 Touch Real-Time PCR Detection System: 50°C for 2 min, 95°C for 2min, and 95°C for 1 sec followed by 60°C for 30 sec repeated for 40 cycles. qPCR primers are shown below.

Analysis

Two experimental (technical) replicates were included in each QPCR run and delta ct was calculated for each sample using the mean of the gene of interest’s technical replicates and the house keeping gene’s technical replicates. Delta-delta ct was calculated for each sample by subtracting the median dct value for the corresponding untreated samples. Three independant experiments (biological replicates) were performed in total. The Student’s t-test was used to compare the p-53 wild type and p-53 null group’s delta-delta ct values in each treatment concentration and for both genes.

Primers

primers <- data.frame(
    name=c(
        "ß-actin Fwd",
        "ß-actin Rev",
        "miR34a HG_F",
        "miR34a HG_R",
        "miR34a_asF1",
        "miR34a_asR1"
        
    ),
    sequence=c(
        "AGGTCATCACCATTGGCAATGAG",
        "CTTTCGGGATGTCCACGTCA",
        "TCTGCTCCAGTGGCTGATGAGAAA",
        "GTTCACTGGCCTCAAAGTTGGCAT",
        "AGCGGCATCTCCTCCACCTGAAA",
        "TTGCCTCGTGAGTCCAAGGAGAAT"
        
    )
)
primers

Results

data <- getData("Figure 2b")

#calculate dct
groups <- c("Condition", "Treatment", "Biological Replicate", "Cell line", "gene")
gois <- c("lncTAM34a", "miR34a HG")

data <- technicalMeans(data, groups) %>%
  dct(., gois, Actin, groups)

#calculate ddct
groups[groups == "gene"] <- "GOI"
logical <- tibble(
  Treatment = data$Treatment == 0
)

data <- data %>%
  ddct(., logical, groups) %>%
  folds(.)

#calculate stats
stats <- calcStats(data, Condition, "HCT116 p53-wt", groups) %>%
  pFormat(.)

#setup plotting variables
stats <- stats %>%
  group_by(GOI, Treatment) %>%
  mutate(
    max = max(CI95h),
    x = case_when(
      Treatment ==   0 ~ 1 + 0.25,
      Treatment == 100 ~ 2 + 0.25,
      Treatment == 200 ~ 3 + 0.25,
      Treatment == 500 ~ 4 + 0.25
    ),
    xend = case_when(
      Treatment ==   0 ~ 1 - 0.25,
      Treatment == 100 ~ 2 - 0.25,
      Treatment == 200 ~ 3 - 0.25,
      Treatment == 500 ~ 4 - 0.25
    )
  ) %>%
  ungroup()

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(x = Treatment, y = log2fold, fill = Condition),
    colour = NA,
    binwidth = 1/10,
    #alpha = 0.5,
    position = position_dodge(width = 0.9), 
    binaxis='y', 
    stackdir='center',
    dotsize = 3,
    show.legend = TRUE
  ) +
  facet_grid(. ~ GOI) +
  scale_y_continuous("\u0394\u0394Ct", breaks = c(-1,0,1,2,3,4,5))+
  labs(x = "Doxorubicin (ng/ml)")

markdown <- p + 
  geom_point(
    data = stats,
    aes(x = Treatment, y = mean, group = Condition), 
    position = position_dodge(width = 0.9),
    colour = "gray20",
    size = 6,
    shape = 95
  ) +
  geom_linerange(
    data = stats,
    aes(x = Treatment, ymin = CI95l, ymax = CI95h, group = Condition),
    colour = "gray20",
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_label(
    data = stats,
    aes(x = Treatment, y = max + 0.28, label = pFormat),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 5,
    family = "Arial Unicode MS",
    position = position_dodge(width = 0.9)
  ) +
  geom_segment(
    data = stats,
    aes(x = x, y = max + 0.1, xend = xend, yend = max + 0.1),
    colour="grey43",
    show.legend = FALSE
  )  +
  guides(
    fill = guide_legend(
      title = "Cell line", 
      override.aes = list(size = 6)
    )
  )

plotRmarkdown(markdown)

pdf <- p + 
  geom_point(
    data = stats,
    aes(x = Treatment, y = mean, group = Condition), 
    position = position_dodge(width = 0.9),
    colour = "grey30",
    shape = 95,
    size = 4
  ) +
  geom_linerange(
    data = stats,
    aes(x = Treatment, ymin = CI95l, ymax = CI95h, group = Condition),
    colour = "grey30",
    position = position_dodge(width = 0.9),
    size = 0.3,
    show.legend = FALSE
  ) +
  geom_label(
    data = stats,
    aes(
      x = Treatment, 
      y = max + 0.4, 
      label = pFormat
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    colour = "grey30",
    size = 2.5,
    family = "Arial Unicode MS",
    position = position_dodge(width = 0.9)
  ) +
  geom_segment(
    data = stats,
    aes(x = x, y = max + 0.1, xend = xend, yend = max + 0.1),
    colour="grey30",
    size = 0.3,
    show.legend = FALSE
  )   +
  guides(
    fill = guide_legend(
      title = "Cell line", 
      override.aes = list(size = 4)
    )
  )

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 2b"][[1]])
# ggsave(
#   plot = figure,
#    filename = path,
#   device = cairo_pdf,
#   height = 60,
#   width = 115,
#   units = "mm"
# )

Conclusions

We observed increasing miR34a HG and asRNA expression with increasing doses of doxorubicin indicating that these two transcripts are co-regulated by p53. Although p53 wild type cells showed a 25-fold increase over untreated cells in miR34a AS expression at the highest tested doxorubicin concentration, this effect was largely abrogated, 11 fold over untreated, in p53 null cells, indicating that p53 is a major regulator of miR34a asRNA expression.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] tidyselect_0.2.3       ggfittext_0.5.0        XML_3.98-1.9          
## [43] crayon_1.3.4           withr_2.1.1.9000       nlme_3.1-131          
## [46] jsonlite_1.5           magrittr_1.5           scales_0.5.0.9000     
## [49] cli_1.0.0.9001         stringi_1.1.6          reshape2_1.4.3        
## [52] xml2_1.2.0             RColorBrewer_1.1-2     rematch2_2.0.1        
## [55] tools_3.4.3            glue_1.2.0             hms_0.4.1             
## [58] parallel_3.4.3         survival_2.41-3        yaml_2.1.16           
## [61] colorspace_1.3-2       rvest_0.3.2            bindr_0.1             
## [64] haven_1.1.1

Introduction

It is likely, due to the head-to head orientation of miR34a HG and asRNA, that transcription initiation may be activated from a single promoter in a bi-directional manner. To investigate whether miR34a HG and asRNA are transcribed from the same promoter as convergent transcripts, we cloned the miR34a HG promoter, including the p53 binding site, into a luciferase/renilla dual reporter vector which we hereafter refer to as p1. The p1 sequence was previously published in Raver-Shapira, N., et al., Transcriptional activation of miR-34a contributes to p53-mediated apoptosis. Mol Cell, 2007. 26(5): p. 731-43.

Methods

Cell culture, transfection and luminescence quantification

All cell lines were cultured at 5% CO2 and 37° C with HEK293T cells cultured in DMEM high glucose (Hyclone) and HCT116 cells in McCoy’s 5a (Life Technologies). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. Cells were plated at 10,000 cells per well in a 96-well plate with a white bottom and cultured overnight. The following day cells were co-transfected with 10ng of empty, p1, or p2 plasmid and GFP using the standard lipofektamine 2000 (Life Technologies) protocol. The expression of GFP and luminescence was measured 24 h post transfection by using the Dual-Glo Luciferase Assay System (Promega) and detected by the GloMax-Multi+ Detection System (Promega). The expression of luminescence was normalized to GFP.

Analysis

Three independant experiments (biological replicates) were performed in total. The Student’s t-test was used to compare the empty vs p1 group’s log2 activity for both genes.

Constructs

ps <- data.frame(
  name = c("p1"),
  sequence = c("GCG CCC TGC CTG GCC CCC ACC TGG TCC TCT TTC CTT TTC AGG TGG AGG AGA TGC CGC TGT
              CCC GTC GGT CTG GGG ACA GCC CAG CTC CCC GGA TCC CGG GCT GGA GAG ACG CGT CGC GGC   
              CCC GGG GCC TGG TGG CAC GAG CAG GAA GGA GGA CCC GGC GGC GGG CTC TGC CTG GGC TTG 
              CCT GGG CTT GTT CCG AGC CGG GCT GCT TCT CGG TGA CCA CGC AGA TCG GGG GCA TTT GGA
              GAT TTT GCG GGA GTC CTG CAG CCA AGC TCC GGG GCA GGA GAG GCC TGG AAG CCT GCA CTA
              CCT GCT C"
  )
)
ps

Primers

primers <- data.frame(
    name=c(
        "Luc_set_II_F",
        "Luc_set_II_R",
        "Renilla_pBiDir_F1",
        "Renilla_pBiDir_R1",
        "B-actin_F",
        "B-actin_R"
    ),
    sequence=c(
        "AAG ATT CAA AGT GCG CTG CTG",
        "TTG CCT GAT ACC TGG CAG ATG",
        "TAA CGC GGC CTC TTC TTA TTT",
        "GAT TTG CCT GAT TTG CCC ATA",
        "AGG TCA TCA CCA TTG GCA ATG AG",
        "CTT TGC GGA TGT CCA CGT CA"
    )
)
primers

Results

*For a schematic representation of the P1 construct please see Supplementary Figure 3a-b.

data <- getData("Figure 2c")

.calcP <- function(`Cell line`, gene, logValue, data) {
  bool1 <- pull(data, `Cell line`) == `Cell line`
  bool2 <- pull(data, gene) == gene
  bool3 <- pull(data, construct) == "Empty"
  bool <- bool1 & bool2 & bool3
  t.test(logValue, pull(data, logValue)[bool])$p.value
}

data <- data %>%
    mutate(logValue = log2(value))

stats <- data %>%
  group_by(`construct`, `Cell line`, gene) %>%
  summarize(
    n = n(),
    mean = mean(logValue),
    CI95l = t.test(logValue)$conf.int[1],
    CI95h = t.test(logValue)$conf.int[2],
    pValue = .calcP(`Cell line`, gene, logValue, data)
  ) %>%
  ungroup() %>%
  pFormat(.)

#vertical lines and labels in plot
ys <- stats %>%
  group_by(`Cell line`) %>%
  summarize(max = max(CI95h)) 

vl <- stats %>%
  group_by(`Cell line`, gene) %>%
  summarize(
    pValue = as.numeric(min(pValue))
  ) %>%
  ungroup() %>%
  pFormat(.) %>%
  mutate(
    x = case_when(
      gene == "Renilla" ~ 1 + 0.25,
      gene == "Luciferase" ~ 1 - 0.25
    ),
    xend = case_when(
      gene == "Renilla" ~ 2 + 0.25,
      gene == "Luciferase" ~ 2 - 0.25
    ),
    pFormat = pFormat
  ) %>%
  left_join(ys) %>%
  mutate(max = max + c(0.25, 1.2, 0.25, 1.2))

#fix gene labels
data <- data %>%
  mutate(plotGene = case_when(
    gene == "Luciferase" ~ "  Luciferase\n(miR34a HG)",
    gene == "Renilla"    ~ "    Renilla\n(lncTAM34a)",
    TRUE ~ "somehow missed a gene"
  ))

stats <- stats %>%
  mutate(plotGene = case_when(
    gene == "Luciferase" ~ "  Luciferase\n(miR34a HG)",
    gene == "Renilla"    ~ "    Renilla\n(lncTAM34a)",
    TRUE ~ "somehow missed a gene"
  ))

vl <- vl %>%
  mutate(plotGene = case_when(
    gene == "Luciferase" ~ "  Luciferase\n(miR34a HG)",
    gene == "Renilla"    ~ "    Renilla\n(lncTAM34a)",
    TRUE ~ "somehow missed a gene"
  ))

data <- filter(data, construct %in% c("p1", "Empty"))
stats <- filter(stats, construct %in% c("p1", "Empty"))

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = construct,
      y = logValue,
      fill = plotGene
    ),
    binwidth = 1/10,
    #alpha = 0.5,
    colour = NA,
    position = position_dodge(width = 0.9), 
    binaxis = 'y', 
    stackdir = 'center',
    dotsize = 3,
    show.legend = TRUE
  ) +
  facet_grid(. ~ `Cell line`) +
  labs(
    x = "Construct",
    y = "log2 activity (a.u.)"
  )

markdown <- p +
  geom_linerange(
    data = stats,
    aes(
      x = construct,
      ymin = CI95l,
      ymax = CI95h,
      group = plotGene
    ),
    colour = "grey20",
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = construct,
      y = mean,
      group = plotGene
    ),
    position = position_dodge(width = 0.9),
    show.legend = FALSE,
    colour = "gray20",
    size = 6,
    shape = 95
  ) +
  geom_label(
    data = vl,
    aes(
      x = 1.5, 
      y = max + 0.25, 
      label = pFormat, 
      group = gene
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 5,
    family = "Arial Unicode MS",
    position = position_dodge(width = 0.9)
  ) +
  geom_segment(
    data = vl,
    aes(
      x = x,
      y = max,
      xend = xend,
      yend = max
    ),
    colour = "grey43",
    show.legend = FALSE
  ) +
  guides(fill = guide_legend(
    title = "Gene", 
    override.aes = list(size = 6)
  ))

plotRmarkdown(markdown)

pdf <- p +
  geom_linerange(
    data = stats,
    aes(
      x = construct,
      ymin = CI95l,
      ymax = CI95h,
      group = plotGene
    ),
    colour = "grey30",
    size = 0.3,
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = construct,
      y = mean,
      group = plotGene
    ),
    position = position_dodge(width = 0.9),
    show.legend = FALSE,
    colour = "grey30",
    size = 4,
    shape = 95
  ) +
  geom_label(
    data = vl,
    aes(
      x = 1.5, 
      y = max + 0.4, 
      label = pFormat, 
      group = gene
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    colour = "grey30",
    size = 2.5,
    family = "Arial Unicode MS",
    position = position_dodge(width = 0.9)
  ) +
  geom_segment(
    data = vl,
    aes(
      x = x,
      y = max,
      xend = xend,
      yend = max
    ),
    colour = "grey30",
    size = 0.3,
    show.legend = FALSE
  ) +
  guides(fill = guide_legend(
    title = "Gene", 
    override.aes = list(size = 4)
  ))

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 2c"][[1]])
# ggsave(
#    plot = figure,
#    filename = path,
#    device = cairo_pdf,
#    height = 60,
#    width = 68,
#    units = "mm"
# )

Conclusion

Upon transfection of this construct in two separate cell lines we observed increases in both luciferase and renilla indicating that miR34a HG and asRNA expression is regulated by a single promoter contained within the p1 construct.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] tidyselect_0.2.3       ggfittext_0.5.0        XML_3.98-1.9          
## [43] crayon_1.3.4           withr_2.1.1.9000       nlme_3.1-131          
## [46] jsonlite_1.5           magrittr_1.5           scales_0.5.0.9000     
## [49] cli_1.0.0.9001         stringi_1.1.6          reshape2_1.4.3        
## [52] xml2_1.2.0             RColorBrewer_1.1-2     rematch2_2.0.1        
## [55] tools_3.4.3            glue_1.2.0             hms_0.4.1             
## [58] parallel_3.4.3         survival_2.41-3        yaml_2.1.16           
## [61] colorspace_1.3-2       rvest_0.3.2            bindr_0.1             
## [64] haven_1.1.1

Introduction

Our previous results showed that miR34a HG and asRNA are simultaneously up-regulated in response to doxorubicin treatment leading us to consider the possibility that miR34a asRNA may be necessary for an appropriate miR34a response to apoptotic signaling. To explore this possibility we again utilized the p1 construct, co-transfecting it with an shRNA targeting renilla, and treating HCT116 cells with increasing levels of doxorubicin.

Methods

Cell culture, treatment, and QPCR

We utilized the P1 construct where the overlapping region of miR34a HG and miR34a AS is cloned with luciferase downstream of miR34a HG and renilla downstream of miR34a AS. The p1 sequence was previously published in Raver-Shapira, N., et al., Transcriptional activation of miR-34a contributes to p53-mediated apoptosis. Mol Cell, 2007. 26(5): p. 731-43. All cell lines were cultured at 5% CO2 and 37° C with HCT116 cells cultured in McCoy’s 5a (Life Technologies). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. 2.5x10^5 HCT116wt cells were seeded in a 12-well plate. After 24hrs these were co-transfected with the P1 construct (25ng) and shRenilla2.1 (250ng) using lipo2000 (Life Technologies) standard protocol. 24hrs post-transfection, doxorubicin treatment was initiated using 0, 300, or 500ng/ml. 24hrs post-treatment, RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. QPCR was carried out using KAPA 2G SYBRGreen (Kapa Biosystems) using the Applied Biosystems 7900HT machine with the cycling conditions: 95 °C for 3 min, 95 °C for 3 s, 60 °C for 30 s. All constructs are shown below.

Analysis

Two experimental (technical) replicates were included in each QPCR run and delta ct was calculated for each sample using the mean of the gene of interest’s technical replicates and the house keeping gene’s technical replicates. delta-delta ct was calculated for each sample by subtracting the dct value for the corresponding shControl samples at treatment concentration 0. Three independant experiments (biological replicates) were performed in total. The Student’s t-test was used to compare the shControl vs shRenilla group’s delta-delta ct values.

Constructs

p1 <- data.frame(
    name="p1",
    sequence="GCG CCC TGC CTG GCC CCC ACC TGG TCC TCT TTC CTT TTC AGG TGG AGG AGA TGC CGC TGT
              CCC GTC GGT CTG GGG ACA GCC CAG CTC CCC GGA TCC CGG GCT GGA GAG ACG CGT CGC GGC   
              CCC GGG GCC TGG TGG CAC GAG CAG GAA GGA GGA CCC GGC GGC GGG CTC TGC CTG GGC TTG 
              CCT GGG CTT GTT CCG AGC CGG GCT GCT TCT CGG TGA CCA CGC AGA TCG GGG GCA TTT GGA
              GAT TTT GCG GGA GTC CTG CAG CCA AGC TCC GGG GCA GGA GAG GCC TGG AAG CCT GCA CTA
              CCT GCT C"
)
p1

shRNAs

shRNAs <- data.frame(
    name=c(
        "shRenilla2.1"
    ),
    sequence=c(
        "TAA CGG GAT TTC ACG AGG C"
    )
)
shRNAs

Primers

primers <- data.frame(
    name=c(
        "Luc_set_II_F",
        "Luc_set_II_R",
        "Renilla_pBiDir_F1",
        "Renilla_pBiDir_R1",
        "B-actin_F",
        "B-actin_R"
    ),
    sequence=c(
        "AAG ATT CAA AGT GCG CTG CTG",
        "TTG CCT GAT ACC TGG CAG ATG",
        "TAA CGC GGC CTC TTC TTA TTT",
        "GAT TTG CCT GAT TTG CCC ATA",
        "AGG TCA TCA CCA TTG GCA ATG AG",
        "CTT TGC GGA TGT CCA CGT CA"
    )
)
primers

Results

See Supplementary Figure 3a-b for a schematic representation of the P1 construct.

data <- getData("Figure 2d")

#Luciferase = miR34a HG
#Renilla = miR34a AS

#calculate dct
groups <- c("treatment", "Biological Replicate", "shRNA", "gene")
gois <- c("Luciferase", "Renilla")

data <- data %>%
  technicalMeans(., groups) %>%
  dct(., gois, Actin, groups)

#calculate ddct
groups[groups == "gene"] <- "GOI"
logical <- tibble(
  `shRNA` = pull(data, shRNA) == "shRNA Control",
  treatment = pull(data, treatment) == 0
)

data <- data %>%
  ddct(., logical, groups, med = FALSE) %>%
  folds(.)

#calculate stats
stats <- calcStats(data, shRNA, "shRNA Control", groups) %>%
  pFormat(.)

#prepare for plotting
data <- data %>%
  mutate(plotGOI = case_when(
    GOI == "Luciferase" ~ " Luciferase\n(miR34a HG)",
    GOI == "Renilla" ~ "Renilla\n(lncTAM34a)"
  )) %>%
  mutate(plotGOI = factor(
    plotGOI, 
    levels = c("Renilla\n(lncTAM34a)", " Luciferase\n(miR34a HG)")
  ))

stats <- stats %>%
  mutate(plotGOI = case_when(
    GOI == "Luciferase" ~ " Luciferase\n(miR34a HG)",
    GOI == "Renilla" ~ "Renilla\n(lncTAM34a)"
  )) %>%
  mutate(plotGOI = factor(
    plotGOI, 
    levels = c("Renilla\n(lncTAM34a)", " Luciferase\n(miR34a HG)")
  )) %>%
  group_by(GOI, treatment) %>%
  mutate(
    height = if_else(treatment == 0, 0.15, max(CI95h, na.rm = TRUE)),
    x = case_when(
      treatment ==   0 ~ 1 + 0.25,
      treatment == 300 ~ 2 + 0.25,
      treatment == 500 ~ 3 + 0.25
    ),
    xend = case_when(
      treatment ==   0 ~ 1 - 0.25,
      treatment == 300 ~ 2 - 0.25,
      treatment == 500 ~ 3 - 0.25
    )
  ) %>%
  ungroup()

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = treatment,
      y = log2fold,
      fill = shRNA
    ),
    binwidth = 1/10,
    #alpha = 0.5,
    colour = NA,
    position = position_dodge(width = 0.9), 
    binaxis = 'y', 
    stackdir = 'center',
    dotsize = 3,
    show.legend = TRUE
  ) +
  facet_grid(. ~ plotGOI, scale = "free") +
  scale_y_continuous("\u0394\u0394Ct") +
  labs(
    x = "Doxorubicin (ng/ml)",
    title = "HCT116"
  )

markdown <- p + 
  geom_linerange(
    data = stats,
    aes(
      x = treatment, 
      ymin = CI95l, 
      ymax = CI95h, 
      group = shRNA
    ),
    colour = "gray20",
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = treatment, 
      y = mean, 
      group = shRNA
    ),
    colour = "gray20",
    position = position_dodge(width = 0.9),
    shape = 95,
    size = 6,
    show.legend = FALSE
  ) +
  geom_label(
    data = stats,
    aes(
      x = treatment, 
      y = height + 0.5,
      label = pFormat
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 5,
    family = "Arial Unicode MS"
  ) +
  geom_segment(
    data = stats,
    aes(x = x, y = height + 0.25, xend = xend, yend = height + 0.25),
    colour="grey43",
    show.legend = FALSE
  )  +
  guides(
    fill = guide_legend(
      title = "Gene", 
      override.aes = list(size = 6)
    )
  )

plotRmarkdown(markdown)

pdf <- p + 
  geom_linerange(
    data = stats,
    aes(
      x = treatment, 
      ymin = CI95l, 
      ymax = CI95h, 
      group = shRNA
    ),
    colour = "grey30",
    position = position_dodge(width = 0.9),
    size = 0.3,
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = treatment, 
      y = mean, 
      group = shRNA
    ),
    colour = "grey30",
    position = position_dodge(width = 0.9),
    shape = 95,
    size = 4,
    show.legend = FALSE
  ) +
  geom_label(
    data = stats,
    aes(
      x = treatment, 
      y = height + 0.6,
      label = pFormat
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    colour = "grey30",
    size = 2.5,
    family = "Arial Unicode MS"
  ) +
  geom_segment(
    data = stats,
    aes(x = x, y = height + 0.25, xend = xend, yend = height + 0.25),
    colour="grey30",
    size = 0.3,
    show.legend = FALSE
  )  +
  guides(fill = guide_legend(
    title = "Gene",
    override.aes = list(size = 4)
  ))

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 2d"][[1]])
# ggsave(
#    plot = figure,
#    filename = path,
#    device = cairo_pdf,
#    height = 60,
#    width = 100,
#    units = "mm"
# )

Conclusion

Levels of luciferase were abrogated in the shRenilla-transfected cells compared to control shRNA-transfected cells. Collectively, these results indicate that miR34a asRNA positively regulates levels of miR34a HG and is crucial for a proper miR34a response to apoptotic stimuli.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] tidyselect_0.2.3       ggfittext_0.5.0        XML_3.98-1.9          
## [43] crayon_1.3.4           withr_2.1.1.9000       nlme_3.1-131          
## [46] jsonlite_1.5           magrittr_1.5           scales_0.5.0.9000     
## [49] cli_1.0.0.9001         stringi_1.1.6          reshape2_1.4.3        
## [52] xml2_1.2.0             RColorBrewer_1.1-2     rematch2_2.0.1        
## [55] tools_3.4.3            glue_1.2.0             hms_0.4.1             
## [58] parallel_3.4.3         survival_2.41-3        yaml_2.1.16           
## [61] colorspace_1.3-2       rvest_0.3.2            bindr_0.1             
## [64] haven_1.1.1

Introduction

We hypothesized that miR34a asRNA may function to mediate the levels of its sense gene, miR34a. To this end, we stably over-expressed miR34a asRNA, using lentiviral transduction, in 3 different p53-null cell lines and monitored the effects on miR34a.

Methods

miR34a asRNA cloning

RNA was harvested from U2OS cells using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. PCR was run using the miR34aAS_cloning_F4 and miR34aAS_cloning_Ex3_1 primers for 40 cycles at annealing temperature 58˚C and run on an agarose gel. Bands were excised and extracted using the QIAquick gel extraction kit (Qiagen). PCR products were cloned using the Strata Clone kit (Aligent) and sequenced (results below). The insert for the lentiviral construct (p-HIV-I2AG-U6) was subsequently sub-cloned from the sequenced plasmid using primers with PacI and NheI cleavage sites. The cloned transcript largley mirrors the GENCODE annotated transcript.

Cloning primers

primers <- data.frame(
    name=c(
        "miR34aAS_cloning_F4",
        "miR34aAS_cloning_Ex3_1"
    ),
    sequence=c(
        "ACGCGTCTCTCCAGCCCGGGAT",
        "AATGATGGCCGCAACTAATGACGG"
    )
)
primers

Sequencing results

construct <- data.frame(
    name="miR34a_asRNA",
    sequence="ACG CGT CTC TCC AGC CCG GGA TCC GGG GAG CTG GGC TGT CCC CAG ACC GAC GGG ACA GCG GCA
                TCT CCT CCA CCT GAA AAG GAA AGA GGA CCA GTT TGC AGG ACT CCG AAC TGG GCC CGC GAG ATC 
                TCC ACC TGC GCA AAA CGA AAG GGC GGA TTC TCC TTG GAC TCA CGA GGC AAC CGC TCC CCG GGG 
                TGA GAA CGG GGG ACT CAT TCC TCC GGC ACT GGG AGA AGA CGA TTC TTC AGG AGG AGG ACA GGG 
                AAG CGA ATG CTA CCC AGA TGT CTC AGT ATA CTG GCT CGC GGC ACA TCG GGC AAA TGA ACC TAT 
                CAG ATA ACA ACG GCA GAT CAG ATG CCT GAG CAT TCA GAA GCA ACA GCT GTG GAG CCC CCG TGG 
                GTT CAG AAG GCC TGG TTC CCG TCT CCA GAA GCC TGG CTC TCC TCC CTC CTG GGC CCA CTA CTT 
                TGG CTT CTT GTT CCT ACG TAC AAG GAG TTG CGA AGA AGG CAA CTC TTC CCC TCC CTG AAG CCA
                AAG GAA TGA AAC AGA CTA GGG CGG GAG AGG TGG CCA TCC GTC ATT AGT TGC GGC CAT CA"
)
construct

Lentivirus production and infection

HEK293T cells were transfected with viral and expression constructs using Lipofectamine 2000 (Life Technologies), after which viral supernatants were harvested 48 and 72 hours post-transfection. Viral particles were concentrated using PEG-IT solution (Systems Biosciences) according to the manufacturer’s recommendations. HEK293T cells were used for virus titration and GFP expression was evaluated 72hrs post-infection via flow cytometry after which TU/ml was calculated. Cell lines were subsequently infected with the lentiviral particals using a multuplicity of infection = 1. After 48 hours, cells were placed under mycophenolic acid selection and the selection process was monitored using GFP expression via flow cytometry. When GFP expression was present in >90% of the cells, cells were frozen for downstream analysis.

QPCR

RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. QPCR was carried out using KAPA 2G SYBRGreen (Kapa Biosystems) using the Applied Biosystems 7900HT machine with the cycling conditions: 95 °C for 3 min, 95 °C for 3 s, 60 °C for 30 s. QPCR for miRNA expression analysis was performed according to the protocol for the TaqMan microRNA Assay kit (Life Technologies) with primer/probe sets purchased from the same company (TaqMan® MicroRNA Assay, hsa-miR-34a, human and Control miRNA Assay, RNU48, human) and the same cycling scheme as above. QPCR primers are shown below.

Analysis

Two experimental (technical) replicates were included in each QPCR run and delta ct was calculated for each sample using the mean of the gene of interest’s technical replicates and the house keeping gene’s technical replicates. delta-delta ct was calculated for each sample by subtracting the median dct value for the corresponding mock samples. Five independant experiments were performed in total. The Student’s t-test was used to compare the mock vs miR34a asRNA overexpressing group’s delta-delta ct values for both genes.

QPCR primers

primers <- data.frame(
    name=c(
        "ß-actin Fwd",
        "ß-actin Rev",
        "miR34a_asF1",
        "miR34a_asR1"
        
    ),
    sequence=c(
        "AGGTCATCACCATTGGCAATGAG",
        "CTTTCGGGATGTCCACGTCA",
        "AGCGGCATCTCCTCCACCTGAAA",
        "TTGCCTCGTGAGTCCAAGGAGAAT"
        
    )
)
primers

Results

data <- getData("Figure 3a")

#Have to divide the analysis due to the presence of several house 
#keeping genes.
gois1 <- c("lncTAM34a", "Actin")
gois2 <- c("miR34a", "RNU48")
data1 <- filter(data, gene %in% gois1)
data2 <- filter(data, gene %in% gois2)

#calculate dct
groups <- c("Genetic mod", "Biological Replicate", "Cell line", "gene")

data1 <- data1 %>%
  technicalMeans(., groups) %>%
  dct(., gois1, Actin, groups, F) %>%
  mutate(GOI = as.character(GOI))

data2 <- data2 %>%
  technicalMeans(., groups) %>%
  dct(., gois2, RNU48, groups, F) %>%
  mutate(GOI = as.character(GOI))

data <- bind_rows(data1, data2) %>%
  mutate(GOI = parse_factor(GOI, levels = c("lncTAM34a", "miR34a")))

#calculate ddct
groups[groups == "gene"] <- "GOI"
logical <- tibble(
  `Biological Replicate` = pull(data, `Biological Replicate`) == 2,
  `Genetic mod` = pull(data, `Genetic mod`) == "mock"
)

data <- data %>%
  ddct(., logical, groups) %>%
  folds(.)

#calculate stats
stats <- calcStats(data, `Genetic mod`, "mock", groups) %>%
  pFormat(.)

#format for plotting
lev <- c("mock", "lncTAM34a", "Skov3\nmock", "Skov3\nlncTAM34a", "Saos2\nmock",   "Saos2\nlncTAM34a")

#data <- data %>%
#  mutate(`Genetic mod` = gsub("miR34a asRNA", "miR34a\nasRNA", `Genetic mod`)) %>%
#  mutate(`Genetic mod` = parse_factor(`Genetic mod`, levels = c("mock", "miR34a\nasRNA")))

#stats <- stats %>%
#  mutate(`Genetic mod` = gsub("miR34a asRNA", "miR34a\nasRNA", `Genetic mod`)) %>%
#  mutate(`Genetic mod` = parse_factor(`Genetic mod`, levels = c("mock", "miR34a\nasRNA")))

#vertical lines and labels in plot
ys <- stats %>%
  group_by(`Cell line`) %>%
  summarize(max = max(CI95h)) 

vl <- stats %>%
  group_by(`Cell line`, GOI) %>%
  summarize(
    pValue = as.numeric(min(pValue))
  ) %>%
  ungroup() %>%
  pFormat(.) %>%
  mutate(
    x = case_when(
      GOI ==    "miR34a" ~ 1 + 0.25,
      GOI == "lncTAM34a" ~ 1 - 0.25
    ),
    xend = case_when(
      GOI ==    "miR34a" ~ 2 + 0.25,
      GOI == "lncTAM34a" ~ 2 - 0.25
    )
  ) %>%
  inner_join(ys) %>%
  mutate(
    max = max + c(0.95, 3.35, 0.65, 3.15, 0.65, 3.15)
  )

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = `Genetic mod`,
      y = log2fold,
      fill = GOI,
      group = interaction(`Genetic mod`, GOI)
    ),
    binwidth = 1/10,
    colour = NA,
    #alpha = 0.5,
    position = position_dodge(width = 0.9), 
    binaxis = 'y', 
    stackdir = 'center',
    dotsize = 6,
    show.legend = TRUE
  ) +
  facet_grid(
    . ~ `Cell line`, 
    scales = "free", 
    space = "free"
  ) +
  scale_y_continuous("\u0394\u0394Ct")+
  labs(x = "Cell line")

markdown <- p +
  geom_linerange(
    data = stats,
    aes(
      x = `Genetic mod`,
      ymin = CI95l,
      ymax = CI95h,
      group = interaction(`Genetic mod`, GOI)
    ),
    position = position_dodge(width = 0.9),
    colour = "gray20",
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = `Genetic mod`,
      y = mean,
      group = interaction(`Genetic mod`, GOI)
    ),
    colour = "gray20",
    shape = 95,
    size = 6,
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_label(
    data = vl,
    aes(
      x = 1.5, 
      y = max + 0.4, 
      label = pFormat, 
      group = interaction(`Cell line`, GOI)
    ),
    fill = "white",
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    position = position_dodge(width = 0.9),
    show.legend = FALSE,
    fontface = "bold",
    family = "Arial Unicode MS",
    size = 3
  ) +
  geom_segment(
    data = vl,
    aes(
      x = x,
      y = max,
      xend = xend,
      yend = max
    ),
    colour = "grey43",
    show.legend = FALSE
  ) +
  guides(
    fill = guide_legend(
      title = "Gene", 
      override.aes = list(size = 6)
    )
  )

plotRmarkdown(markdown)

pdf <- p +
  geom_linerange(
    data = stats,
    aes(
      x = `Genetic mod`,
      ymin = CI95l,
      ymax = CI95h,
      group = interaction(`Genetic mod`, GOI)
    ),
    position = position_dodge(width = 0.9),
    colour = "grey30",
    size = 0.3,
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = `Genetic mod`,
      y = mean,
      group = interaction(`Genetic mod`, GOI)
    ),
    colour = "grey30",
    shape = 95,
    size = 4,
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_label(
    data = vl,
    aes(
      x = 1.5, 
      y = max + 1, 
      label = pFormat, 
      group = interaction(`Cell line`, GOI)
    ),
    fill = "white",
    colour = "grey30",
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    position = position_dodge(width = 0.9),
    show.legend = FALSE,
    fontface = "bold",
    family = "Arial Unicode MS",
    size = 2.5
  ) +
  geom_segment(
    data = vl,
    aes(
      x = x,
      y = max,
      xend = xend,
      yend = max
    ),
    colour = "grey30",
    size = 0.3,
    show.legend = FALSE
  ) +
  guides(
    fill = guide_legend(
      title = "Gene", 
      override.aes = list(size = 4)
    )
  )

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 3a"][[1]])
# ggsave(
#    plot = figure,
#    filename = path,
#    device = cairo_pdf,
#    height = 60,
#    width = 83,
#    units = "mm"
# )

Conclusion

miR34a asRNA over-expression causes a concominant up-regulation of miR34a in all 3 cell lines. This indicates that miR34a asRNA concordantly regulates miR34a.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] tidyselect_0.2.3       ggfittext_0.5.0        XML_3.98-1.9          
## [43] crayon_1.3.4           withr_2.1.1.9000       nlme_3.1-131          
## [46] jsonlite_1.5           magrittr_1.5           scales_0.5.0.9000     
## [49] cli_1.0.0.9001         stringi_1.1.6          reshape2_1.4.3        
## [52] xml2_1.2.0             RColorBrewer_1.1-2     rematch2_2.0.1        
## [55] tools_3.4.3            glue_1.2.0             hms_0.4.1             
## [58] parallel_3.4.3         survival_2.41-3        yaml_2.1.16           
## [61] colorspace_1.3-2       rvest_0.3.2            bindr_0.1             
## [64] haven_1.1.1

Introduction

miR34a is known to regulate cell cycle and, specifically, induce G1-phase arrest. Due to the increase observed in miR34a expression levels in cell lines over-expressing miR34a asRNA, we investigated if the cell cycle was perturbed in these cells.

Methods

Cell culture and cell cycle analysis

All cell lines were cultured at 5% CO2 and 37° C with Skov3 and PC3 cells in RPMI (Hyclone) and 2 mM L-glutamine. All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. 1x10^5 cells per well in a 6-well plate and harvested after 24hrs. Cells were washed in PBS and fixed in 4% PFA at room temperature overnight. PFA was removed, and cells were resuspended in 95% EtOH. The samples were then rehydrated in distilled water, stained with DAPI and analyzed by flow cytometry on a LSRII (BD Biosciences) machine.

Analysis

The cell cycle phases ModFit software (Verity Software House) was used to quantify the percentage of cells in each cell cycle phase. Percentages were converted to log2 fractions for each sample. Student’s t-test was used for statistical analysis comparing the mock vs miR34a asRNA overespression conditions using five (PC3) or three (Skov3) independant experiments.

Results

data <- getData("Figure 3b")

#convert the value (precent) to log2(fraction)
data <- mutate(data, value = log2(value / 100))

#qqplot shows that the data is reasonably normal. The shapiro test is also used below. Note that a p < 0.05 indicates non-normality.
#qqnorm(data$value)

#calculate stats
stats <- data %>%
  full_join(
    filter(data, condition == "mock"), 
    by = c("Biological Replicate", "Cell line", "phase"), 
    suffix = c("", ".y")
  ) %>%
  group_by(`Cell line`, condition, phase) %>%
  summarize(
    n = n(),
    mean = (2^mean(.data$value)),
    max = (2^max(.data$value)),
    CI95l = (2^t.test(.data$value)$conf.int[1]),
    CI95h = (2^t.test(.data$value)$conf.int[2]),
    pValue = t.test(.data$value, .data$value.y)$p.value,
    shapiroTest = shapiro.test(value)$p.value
  ) %>%
  ungroup() %>%
  pFormat(.)

#prepare for plotting
#change value back to percent
data <- data %>%
  mutate(value = 2^value)

#calculate start and end for horizontal plot lines
stats <- stats %>%
  group_by(phase, `Cell line`) %>%
  mutate(y = max(c(.data$CI95h, .data$max))) %>%
  ungroup() %>%
  arrange(`Cell line`, phase) %>%
  mutate(
    x = rep(sort(rep(1:3, 2)), 2) - 0.25,
    xend = rep(sort(rep(1:3, 2)), 2) + 0.25
  )

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = phase,
      y = value,
      fill = condition
    ),
    binwidth = 1/100,
    colour = NA,
    #alpha = 0.5,
    position = position_dodge(width = 0.9), 
    binaxis = 'y', 
    stackdir = 'center',
    dotsize = 1,
    show.legend = TRUE
  ) +
  facet_grid(
    . ~ `Cell line`, 
    scale = "free"
  ) +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(
    x = "Cell cycle phase",
    y = "",
    title = "Cell cycle"
  )

markdown <- p +
  geom_linerange(
    data = stats,
    aes(
      x = phase, 
      ymin = CI95l, 
      ymax = CI95h, 
      group = interaction(phase, condition)
    ),
    position = position_dodge(width = 0.9),
    colour = "grey25",
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = phase, 
      y = mean, 
      group = condition
    ), 
    colour = "black",
    position = position_dodge(width = 0.9),
    shape = 95,
    size = 6,
    show.legend = FALSE
  ) +
  geom_segment(
    data = stats,
    aes(
      x = x,
      y = y + 0.01, 
      xend = xend, 
      yend = y + 0.01
    ),
    colour = "grey43",
    show.legend = FALSE
  ) +
  geom_label(
    data = stats,
    aes(
      x = phase, 
      y = y + 0.02, 
      label = pFormat
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 3,
    family = "Arial Unicode MS"
  ) +
  guides(
    fill = guide_legend(
      title = "Condition", 
      override.aes = list(size = 6)
    )
  )

plotRmarkdown(markdown)

pdf <- p +
  geom_linerange(
    data = stats,
    aes(
      x = phase, 
      ymin = CI95l, 
      ymax = CI95h, 
      group = interaction(phase, condition)
    ),
    position = position_dodge(width = 0.9),
    colour = "grey30",
    size = 0.3,
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = phase, 
      y = mean, 
      group = condition
    ), 
    colour = "grey30",
    position = position_dodge(width = 0.9),
    shape = 95,
    size = 4,
    show.legend = FALSE
  ) +
  geom_segment(
    data = stats,
    aes(
      x = x,
      y = y + 0.02, 
      xend = xend, 
      yend = y + 0.02
    ),
    colour = "grey30",
    size = 0.3,
    show.legend = FALSE
  ) +
  geom_label(
    data = stats,
    aes(
      x = phase, 
      y = y + 0.038, 
      label = pFormat
    ),
    colour = "grey30",
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 2.5,
    family = "Arial Unicode MS"
  ) +
  guides(
    fill = guide_legend(
      title = "Condition", 
      override.aes = list(size = 4)
    )
  )

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 3b"][[1]])
# ggsave(
#    plot = figure,
#    filename = path,
#    device = cairo_pdf,
#    height = 60,
#    width = 93,
#    units = "mm"
# )

Conclusion

An increase in the percentage of cells in G1-phase can be seen in cell lines overexpressing miR34a asRNA. In addition, a concominant decrease of cells in G2-phase can be observed.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] bindrcpp_0.2                  gtable_0.2.0                 
##  [3] ggthemes_3.4.0                printr_0.1                   
##  [5] forcats_0.2.0                 stringr_1.2.0                
##  [7] dplyr_0.7.4                   purrr_0.2.4                  
##  [9] readr_1.1.1                   tidyr_0.8.0                  
## [11] tibble_1.4.2                  ggplot2_2.2.1.9000           
## [13] tidyverse_1.2.1               knitr_1.19                   
## [15] miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] broom_0.4.3            compiler_3.4.3         modelr_0.1.1          
## [37] pkgconfig_2.0.1        mnormt_1.5-5           htmltools_0.3.6       
## [40] tidyselect_0.2.3       ggfittext_0.5.0        XML_3.98-1.9          
## [43] crayon_1.3.4           withr_2.1.1.9000       nlme_3.1-131          
## [46] jsonlite_1.5           magrittr_1.5           scales_0.5.0.9000     
## [49] cli_1.0.0.9001         stringi_1.1.6          reshape2_1.4.3        
## [52] xml2_1.2.0             RColorBrewer_1.1-2     rematch2_2.0.1        
## [55] tools_3.4.3            glue_1.2.0             hms_0.4.1             
## [58] parallel_3.4.3         survival_2.41-3        yaml_2.1.16           
## [61] colorspace_1.3-2       rvest_0.3.2            bindr_0.1             
## [64] haven_1.1.1

Introduction

As p53 is a well-known regulator of senescence, growth, and apoptosis and has been shown to specifically important under conditions of cellular stress, such as starvation. Therefore, we investigated whether overexpression of miR34a AS affects growth rate under normal and starvation conditions by measuring changes in confluency over time.

Methods

Cell culture and confluency quantification

10^4 PC3 cells, either miR34a asRNA overexpressing or WT, were seeded in 96 well plates. After attachment (3-4h) media was replaced with either RPMI (Gibco, life technology) (supplemented with 2mM L-glutamine, 50ug/ml Penicillin-Streptomycin and 10% Fetal Calf Serum) or HBSS. Cells were incubated in Spark Multimode Microplate reader for 48 hours at 37°C with 5% CO2 in a humidity chamber. Confluency was measured every hour.

Analysis

Three technical triplicates were included in each experiment and these were independantly normalized for each condition to the 0 time point afterwhich the mean was calculated. The 95% confidence interval was then calculated based on the three independant experiments which were preformed.

Results

data <- getData("Figure 3c")

#normalize to first time point for control (Mock)
.normFun <- function(
  data,
  Confluency,
  Treatment,
  Biological.Replicate,
  Technical.Replicate,
  Cell.line
){
  bool1 <- data$Treatment == Treatment
  bool2 <- data$`Biological Replicate` == Biological.Replicate
  bool3 <- data$`Technical Replicate` == Technical.Replicate
  bool4 <- data$Time == 0
  bool5 <- data$`Cell line` == Cell.line
  bool <- bool1 & bool2 & bool3 & bool4 & bool5
  Confluency / pull(data, Confluency)[bool]
}

data <- data %>%
  group_by(Treatment, `Biological Replicate`, `Technical Replicate`, Time, `Cell line`) %>%
  mutate(
    normConfluency = .normFun(
      data,
      Confluency,
      Treatment,
      `Biological Replicate`,
      `Technical Replicate`,
      `Cell line`
    )
  ) %>%
  ungroup()

#calculate mean for each technical replicate
data <- data %>%
  group_by(Treatment, `Biological Replicate`, Time, `Cell line`) %>%
  summarize(
    techMean = mean(normConfluency)
  ) %>%
  ungroup()

#calculate stats
stats <- data %>%
  filter(Time != 0) %>%
  group_by(`Cell line`, Treatment, Time) %>%
  summarize(
    min = min(techMean),
    max = max(techMean),
    mean = mean(techMean),
    median = median(techMean),
    CI95l = t.test(techMean)$conf.int[1],
    CI95h = t.test(techMean)$conf.int[2]
  )

#linear model
model.data <-  data %>%
  mutate(Time = as.numeric(Time)) %>%
  group_by(Treatment) %>%
  mutate(rank = rank(techMean)) %>%
  do(glance(lm(rank ~ Time +`Cell line`, data = .))) %>%
  ungroup() %>%
  rename(pValue = p.value) %>%
  pFormat(.) %>%
  select(-pValue) %>%
  rename(p.value = pFormat)

as.data.frame(model.data)

growth.lm <- data %>%
  mutate(Time = as.numeric(Time)) %>%
  group_by(Treatment) %>%
  mutate(rank = rank(techMean)) %>%
  do(tidy(lm(rank ~ Time +`Cell line`, data = .))) %>%
  ungroup() %>%
  rename(pValue = p.value) %>%
  pFormat(.) %>%
  select(-pValue) %>%
  rename(p.value = pFormat)

as.data.frame(growth.lm)

p <- ggplot(data = NULL) +
  geom_ribbon(
    data = stats,
    aes(
      x = Time,
      ymin = CI95l, 
      ymax = CI95h, 
      group = `Cell line`,
      fill = `Cell line`
    ), 
    alpha = 0.2,
    show.legend = FALSE
  ) +
  facet_grid(. ~ Treatment) +
  scale_x_discrete(breaks = seq(0, 35, 5)) +
  labs(
    x = "Time (hours)", 
    y = "Normalized % confluency",
    title = "miR34a asRNA's role in growth regulation."
  ) + 
  guides(
    colour = guide_legend(title = "Cell line", override.aes = list(size = 3))
  )

markdown <- p +
  geom_linerange(
    data = stats,
    aes(
      x = Time,
      colour = `Cell line`,
      ymin = min,
      ymax = max
    ), 
    position = position_dodge(width = 1),
    show.legend = FALSE
  ) + 
  geom_point(
    data = stats,
    aes(
      x = Time, 
      y = median,
      group = `Cell line`,
      colour = `Cell line`
    ), 
    position = position_dodge(width = 1),
    show.legend = TRUE
  ) + 
  guides(
    colour = guide_legend(
      title = "Cell line", 
      override.aes = list(size = 3)
    )
  )

plotRmarkdown(markdown)

pdf <- p +
  geom_linerange(
    data = stats,
    aes(
      x = Time,
      colour = `Cell line`,
      ymin = min,
      ymax = max
    ), 
    position = position_dodge(width = 1),
    lwd = 0.2,
    show.legend = FALSE
  ) + 
  geom_point(
    data = stats,
    aes(
      x = Time, 
      y = median,
      group = `Cell line`,
      colour = `Cell line`
    ), 
    size = 0.25,
    position = position_dodge(width = 1),
    show.legend = TRUE
  ) + 
  guides(
    colour = guide_legend(
      title = "Cell line", 
      override.aes = list(size = 2)
    )
  )

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 3c"][[1]])
# ggsave(
#   plot = figure,
#   filename = path,
#   device = cairo_pdf,
#   height = 60,
#   width = 83,
#   units = "mm"
# )

Conclusions

miR34a asRNA overexpression causes a minor decrease in cell growth under normal conditions (RPMI) although under conditions of cellular stress through starvation, this effect is increased dramatically.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] broom_0.4.3                   bindrcpp_0.2                 
##  [3] gtable_0.2.0                  ggthemes_3.4.0               
##  [5] printr_0.1                    forcats_0.2.0                
##  [7] stringr_1.2.0                 dplyr_0.7.4                  
##  [9] purrr_0.2.4                   readr_1.1.1                  
## [11] tidyr_0.8.0                   tibble_1.4.2                 
## [13] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [15] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] survival_2.41-3        yaml_2.1.16            colorspace_1.3-2      
## [61] rvest_0.3.2            bindr_0.1              haven_1.1.1

Introduction

miR34a asRNA over-expression was shown to increase the expression levels of miR34a. We investigated if this increase of miR34a was due to increased levels of transcription at the miR34a promoter.

Methods

Cell culture, QPCR, and ChIP

All cell lines were cultured at 5% CO2 and 37° C with Skov3 and PC3 cells in RPMI (Hyclone) and 2 mM L-glutamine and Saos2 cells cultured in DMEM high modified (Hyclone). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. Cells were plated at 2-3x10^6 cells in a 15cm dish and incubated overnight. Cells were cross-linked in 1% formaldehyde, quenched in glycine (0.125M), resuspended in cell lysis buffer (5mM PIPES, 85mM KCL, 0.5% NP40, protease inhibitor), and sonicated (Bioruptor Sonicator Diagenode) in nucleus lysis buffer (50mM TRIS-HCL pH 8.0, 10mM EDTA, 1% SDS, protease inhibitor). The samples were incubated overnight at 4 °C with the appropriate antibody. Salmon sperm DNA/Protein A–agarose (Upstate/Millipore) was used to pull down the antibody. DNA was eluted in elution buffer (1% SDS, 100 mM NaHCO3), followed by reverse cross-linking, RNase-A and protease-K treatment. The DNA was eluted by using a Qiagen PCR purification kit. The following antibodies were used (4 μg/sample): phospho-polymerase II (abcam: ab5095). QPCR was carried out using KAPA 2G SYBRGreen (Kapa Biosystems) using the Applied Biosystems 7900HT machine with the cycling conditions: 95 °C for 3 min, 95 °C for 3 s, 60 °C for 30 s.

Analysis

QPCR was performed in technical replicate for each sample and the mean of the technical replicates was subsequently used to calculate the fraction of input. The fold change of the miR34a asRNA overexpression samples was then calculated by dividing the values by the values for the corresponding mock samples. The fold values were then log2 transformed and the Student’s t-test was used to test for significant differences under the null hypothesis that the true mean value of log2 transformed fold changes were equal to 0.

Primers

primers <- data.frame(
    name=c(
        "miR34aChIP F1",
        "miR34aChIP R1"
    ),
    sequence=c(
       "AAA GTT TGC AAA GAA GGA GGC GGG",
       "AGG GAA GAA AGA ACT AGC CGA GCA"
    )
)
primers

Results

data <- getData("Figure 3d")

#calculate means of technical replicates
data <- mutate(data, techMean = rowMeans(select(data, Quantity1, Quantity2)))

#calculate fraction of input
.fractionOfInput <- function(data, techMean, Biological.Replicate, condition) {
  bool1 <- data$`Biological Replicate` == Biological.Replicate
  bool2 <- data$condition == condition
  bool3 <- data$sample == "input"
  bool <- bool1 & bool2 & bool3
  techMean / pull(data, techMean)[bool]
}

data <- data %>%
  group_by(`Biological Replicate`, condition, sample) %>%
  mutate(fractionInput = .fractionOfInput(data, techMean, `Biological Replicate`, condition)) %>%
  ungroup() %>%
  filter(sample != "input")

#calculate fold over mock
.foldOverControl <- function(data, fractionInput, Biological.Replicate) {
  bool1 <- data$`Biological Replicate` == 3
  bool2 <- data$condition == "mock"
  bool <- bool1 & bool2
  fractionInput / pull(data, fractionInput)[bool]
}

data <- data %>%
  mutate(
    foldOverControl = .foldOverControl(data, fractionInput, `Biological Replicate`),
    logfoldOverControl= log2(foldOverControl)
  ) %>%
  ungroup()

#calculate stats 
stats <- data %>%
  group_by(condition) %>%
  summarize(
    n = n(),
    mean = mean(logfoldOverControl),
    CI95l =t.test(logfoldOverControl)$conf.int[1],
    CI95h =t.test(logfoldOverControl)$conf.int[2],
    pValue = t.test(logfoldOverControl, mu = 0)$p.value
  ) %>%
  pFormat(.) %>%
  mutate(pFormat = case_when(
    condition == "mock" ~ "", 
    TRUE ~ pFormat
  ))

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = condition, 
      y = logfoldOverControl, 
      fill = condition
    ),
    binwidth = 1/10,
    #alpha = 0.5,
    colour = NA,
    position = position_dodge(width = 0.9), 
    binaxis='y', 
    stackdir='center',
    dotsize = 2,
    show.legend = TRUE
  ) +
  scale_y_continuous("log2(Fold)") +
  labs(
    x = "Cell line",
    y = "log2(Fold polII enrichment)",
    title = "Phosphorylated Poll II ChIP"
  ) +
  guides(fill = FALSE)

markdown <- p +
    geom_linerange(
    data = stats,
    aes(x = condition, ymin = CI95l, ymax = CI95h),
    colour = "gray20",
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(x = condition, y = mean),
    colour = "gray20",
    shape = 95,
    size = 6,
    show.legend = FALSE
  ) +
  geom_segment(
    data = filter(stats, condition == "lncTAM34a"),
    aes(x = 1, y = CI95h + 0.25, xend = 2, yend = CI95h + 0.25),
    colour = "grey43"
  ) +
  geom_label(
    data = stats,
    aes(x = 1.5, y = CI95h + 0.45, label = pFormat),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 5,
    family = "Arial Unicode MS"
  )

plotRmarkdown(markdown)

pdf <- p +
    geom_linerange(
    data = stats,
    aes(x = condition, ymin = CI95l, ymax = CI95h),
    colour = "grey30",
    size = 0.3,
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(x = condition, y = mean),
    colour = "grey30",
    shape = 95,
    size = 4,
    show.legend = FALSE
  ) +
  geom_segment(
    data = filter(stats, condition == "lncTAM34a"),
    aes(x = 1, y = CI95h + 0.25, xend = 2, yend = CI95h + 0.25),
    colour = "grey30",
    size = 0.3
  ) +
  geom_label(
    data = stats,
    aes(x = 1.5, y = CI95h + 0.45, label = pFormat),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    colour = "grey30",
    size = 2.5,
    family = "Arial Unicode MS"
  )

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 3d"][[1]])
# ggsave(
#   plot = figure,
#   filename = path,
#   device = cairo_pdf,
#   height = 60,
#   width = 43,
#   units = "mm",
#   dpi = 10000000
# )

Conclusion

Phosphorylated pol-II was increased at the miR34a promoter in miR34a asRNA over-expressing cell lines indicating that the regulation of miR34a by miR34a asRNA is transcriptional.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] broom_0.4.3                   bindrcpp_0.2                 
##  [3] gtable_0.2.0                  ggthemes_3.4.0               
##  [5] printr_0.1                    forcats_0.2.0                
##  [7] stringr_1.2.0                 dplyr_0.7.4                  
##  [9] purrr_0.2.4                   readr_1.1.1                  
## [11] tidyr_0.8.0                   tibble_1.4.2                 
## [13] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [15] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] survival_2.41-3        yaml_2.1.16            colorspace_1.3-2      
## [61] rvest_0.3.2            bindr_0.1              haven_1.1.1

Introduction

Kaplan-Meier survival curves comparing the effects of TP53-mutated samples, low lncTAM34a expression and low miR34a expression to control samples in 17 cancers from TCGA. Panels showing survival curves based on gene expression include only TP53 wild type patients where RNAseq data exists.

Methods

RNAseq data was downloaded from TCGA and processed as described previously. Briefly, RNAseq data were aligned to the human hg19 assembly and quantified using GENCODE (v19) annotated HTSeq-counts and FPKM normalizations. Expression data from miR34a and lncTAM34a (identified as RP3-510D11.2) were used for further analysis.

Survival analysis was performed on TCGA vital state and follow-up data, downloaded from GDC on 27/10/2017 using the R survival package.

Results

data <- getData("Figure 4a")

#TP53 status independant
#Set up all combinations of variables to calcualte survival probability for
independant <- expand.grid(
  gene = c("lncTAM34a", "miR34a", "TP53"),
  cancer = sort(unique(data$Cancer))
) %>% 
  as_tibble() %>%
  mutate_if(is.factor, as.character) %>%
  group_by(gene, cancer) %>%
  #add data of the corresponding cancer and, for gene expression, filter WT TP53
  mutate(data = map2(cancer, gene, function(c, g, cm = data) {
    if(g == "TP53") {
      filter(cm, Cancer == c)
    } else {
      filter(cm, Cancer == c & TP53 == FALSE)
    }
  })) %>%
  #select the gene to be examined
  mutate(data = pmap(list(gene, data), function(g, d) {
      select(d, g, vitalStatus, FU)
  })) %>%
  #remove samples where gene expression/mutation status is na
  mutate(data = map(data, function(x) filter(x, is.na(x[[1]]) == FALSE))) %>%
  #remove cancers where comparisons without data exist
  mutate(bool = map_lgl(data, function(x) nrow(x) == 0)) %>%
  group_by(cancer) %>%
  filter(!any(bool)) %>%
  #calculate the percentile for the gene's expression for each sample
  mutate(data = map2(gene, data, function(g, d) {
    if(any(grepl("TP53", colnames(d)))) {
      mutate(d, percentile = NA)
    } else {
      mutate(d, percentile = ntile(x = d[[1]], n = 10))
    }
  })) %>%
  #classify 10th percentile and under as "low" and the rest as "others"
  #classify TP53 mutated as low and WT as others
  mutate(data = map(data, function(d) {
    if(any(grepl("TP53", colnames(d)))) {
      mutate(d, class = if_else(TP53, "low", "others"))
    } else {
     mutate(d, class = if_else(percentile == 1, "low", "others")) 
    }
  })) %>%
  #calculate survival probability
  mutate(surv.obj = map(data, ~get_survival(.x))) %>%
  mutate(surv.fit = map2(data, surv.obj, function(x, y) survival_fit(y, x$class))) %>%
  mutate(p.value = map2_dbl(data, surv.obj, function(x, y) survival_p(y, x$class))) %>%
  ungroup()

independant <- filter(independant, cancer == "KIRP")
trash <- pmap(
  list(independant$data, independant$surv.fit, independant$gene, independant$cancer, independant$p.value), 
  function(v, w, x, y, z) {
    plotKM(v, w, x, y, z)
  }
)

Introduction

Correlation analysis between the effects on the 5-year survival probability of TP53-mutated samples, low lncTAM34a expression, and low miR34a expression as indicated. For each variable the 5-year survival probability was compared to the control group (negative values indicate lower survival, positive values indicate higher survival). Spearman correlation coefficients are given on top left of each plot. Each dot indicates one cancer type. Boxplots on the bottom summarize the effects for the parameter on the x-axis, with indication of P values, as calculated using paired Wilcoxon signed rank test. Low expression was defined as TP53 non- mutated samples having expression values in the bottom 10th percentile.

Methods

RNAseq data was downloaded from TCGA and processed as described previously. Briefly, RNAseq data were aligned to the human hg19 assembly and quantified using GENCODE (v19) annotated HTSeq-counts and FPKM normalizations. Expression data from miR34a and lncTAM34a (identified as RP3-510D11.2) were used for further analysis.

Survival analysis was performed on TCGA vital state and follow-up data, downloaded from GDC on 27/10/2017 using the R survival package.

Results

data <- getData("Figure 4b")
correlation_matrix <- filter(data, !is.na(TP53) & !is.na(lncTAM34a))
cancers <- sort(unique(correlation_matrix$Cancer))

#TP53 status independant
#Set up all combinations of variables to calcualte survival probability for
independant <- expand.grid(
  gene = c("lncTAM34a", "miR34a", "TP53"),
  cancer = cancers,
  geneLevel = c("low", "others")
) %>% 
  as_tibble() %>%
  arrange(cancer, gene, geneLevel) %>%
  mutate_if(is.factor, as.character) %>%
  group_by(gene, cancer, geneLevel) %>%
  #add data of the corresponding cancer
  mutate(data = map(cancer, function(c, cm = correlation_matrix) {
    filter(cm, Cancer == c)
  })) %>%
  #select the gene to be examined
  mutate(data = pmap(list(gene, data), function(g, d) {
      select(d, g, vitalStatus, FU)
  })) %>%
  #calculate the percentile for the gene's expression for each sample
  mutate(data = map2(gene, data, function(g, d) {
    if(any(grepl("TP53", colnames(d)))) {
      mutate(d, percentile = NA)
    } else {
      mutate(d, percentile = ntile(x = d[[1]], n = 10))
    }
  })) %>%
  #classify 10th percentile and under as "low" and the rest as "others"
  #classify TP53 mutated as low and WT as others
  mutate(data = map(data, function(d) {
    if(any(grepl("TP53", colnames(d)))) {
      mutate(d, class = if_else(TP53, "low", "others"))
    } else {
     mutate(d, class = if_else(percentile == 1, "low", "others")) 
    }
  })) %>%
  #filter data class to correspond to expected geneLevel
  mutate(data = map2(geneLevel, data, function(gl, d) {
    filter(d, class == gl)
  })) %>%
  #calculate survival probability
  mutate(surv.obj = map(data, ~get_survival(.x))) %>%
  mutate(surv.fit = map(surv.obj, ~survival_fit(.x))) %>%
  mutate(surv.prob = map_dbl(surv.fit, ~calcSurvProb(.x))) %>%
  ungroup() %>%
  select(-data, -surv.fit) %>%
  mutate(type = "independant")


# TP53 status dependent
#Set up all combinations of variables to calcualte survival probability for
dependant <- expand.grid(
  gene = c("lncTAM34a", "miR34a"),
  mutation = c(TRUE, FALSE),
  cancer = cancers,
  geneLevel = c("low", "others")
) %>% 
  as_tibble() %>%
  arrange(cancer, mutation, gene, geneLevel) %>%
  mutate_if(is.factor, as.character) %>%
  group_by(gene, mutation, cancer, geneLevel) %>%
  #add data of the corresponding cancer and TP53 status
  mutate(data = pmap(list(cancer, mutation), function(c, m, cm = correlation_matrix) {
    filter(cm, Cancer == c & TP53 == m)
  })) %>%
  #select the gene to be examined
  mutate(data = pmap(list(gene, data), function(g, d) {
      select(d, g, vitalStatus, FU)
  })) %>%
  #calculate the percentile for the gene's expression for each sample
  mutate(data = map2(gene, data, function(g, d) {
    mutate(d, percentile = ntile(x = d[[1]], n = 10))
  })) %>%
  #classify 10th percentile and under as "low" and the rest as "others"
  mutate(data = map(data, function(d) {
    mutate(d, class = if_else(percentile == 1, "low", "others"))
  })) %>%
  #filter data class to correspond to expected geneLevel
  mutate(data = map2(geneLevel, data, function(gl, d) {
    filter(d, class == gl)
  })) %>%
  #remove comparisons where there is not enough data
  mutate(bool = map_lgl(data, function(d) {
    !nrow(d) <= 1
  })) %>%
  filter(bool) %>%
  #calculate survival probability
  mutate(surv.obj = map(data, ~get_survival(.x))) %>%
  mutate(surv.fit = map(surv.obj, ~survival_fit(.x))) %>%
  mutate(surv.prob = map_dbl(surv.fit, ~calcSurvProb(.x))) %>%
  ungroup() %>%
  select(-(data:surv.fit)) %>%
  mutate(type = "dependent")

#combine the analyses and filter those to be included in the figure
survData <- bind_rows(dependant, independant) %>%
  filter(!is.na(surv.prob)) %>%
  filter((type == "dependent") | (type == "independant" & gene == "TP53")) %>%
  filter(is.na(mutation) | mutation == FALSE) %>%
  group_by(gene, mutation, cancer, type) %>%
  mutate(n = n()) %>%
  filter(n == 2) %>%
  summarize(relSurv = surv.prob[geneLevel == "low"] - surv.prob[geneLevel == "others"]) %>%
  ungroup() %>%
  select(-mutation, -type) %>%
  spread(gene, relSurv)

#Add the colors for the plot to the data
colors <- getCancerColors()
survData <- full_join(survData, colors, by = "cancer")

#add ggplot theme
theme_remove_all <- theme_remove_all()

#plot 1
p1 <- mainPlot(survData, "TP53", "lncTAM34a", tRA = theme_remove_all)
p2 <- hPlot(survData, "TP53", tRA = theme_remove_all)
renderPlots(p1, p2)

#plot 2
p1 <- mainPlot(survData, "lncTAM34a", "miR34a", tRA = theme_remove_all)
p2 <- hPlot(survData, "lncTAM34a", tRA = theme_remove_all)
renderPlots(p1, p2)

#plot 3
p1 <- mainPlot(survData, "miR34a", "TP53", tRA = theme_remove_all)
p2 <- hPlot(survData, "miR34a", tRA = theme_remove_all)
renderPlots(p1, p2)

Introduction

We interogated RNA sequencing data from TCGA to determine in miR34a and miR34a asRNA expression was correlated in both TP53 wild type and mutated samples.

Methods

RNA-Seq data and copy number data were downloaded from TCGA and processed as described previously (Ashouri et al. 2016). Briefly, RNA-Seq data were aligned to the human hg19 assembly and quantified using GENCODE (v19) annotated HTSeq-counts and FPKM normalizations. Expression data from miR34a and miR34 asRNA (identified as RP3-510D11.2) were used for further analysis. Copy number amplitudes for GENCODE genes were determined from segmented copy-number data. Samples that were diploid for miR34 asRNA were identified as those samples that had copy number amplitudes between -0.1 and 0.1.

Somatic mutation data were downloaded from the Genomics Data Commons data portal (GDC) as mutation annotation format (maf) files, called using Mutect2 on 30/10/2017 (v7) (Grossman et al. 2016).

Results

data <- getData("Supplementary Figure 1b")

#normalize expression values
data <- mutate(
    data,
    RP3 = log2(RP3 / max(RP3)),
    miR34a = log2(miR34a / max(miR34a))  
  )

#add BRCA subtypes
data <- data %>%
  mutate(PAM50 = if_else(is.na(PAM50), "", PAM50)) %>%
  mutate(cancer_PAM50 = gsub("BRCA", "BRCA ", paste(cancer, PAM50, sep = "")))

#sort 
data <- arrange(data, cancer_PAM50, TP53, RP3)

#remove infinite data
data <- filter(data, !is.infinite(RP3) & !is.infinite(miR34a))

#diploid only
data <- filter(data, abs(RP3_cna) < 0.1)

##prepare for plotting
data <- data %>%
  mutate(TP53 = if_else(TP53 == 0, "TP53 wt", "TP53 mut")) %>%
  select(TCGA_cancer_id, TP53, cancer_PAM50, RP3, miR34a) %>%
  gather(gene, value, -(TCGA_cancer_id:cancer_PAM50)) %>%
  mutate(gene = if_else(gene == "RP3", "lncTAM34a", "miR34a"))

#remove cancers that don't have at least 5 patients in both TP53 wt and mutated
data <- data %>%
  group_by(cancer_PAM50, TP53, gene) %>%
  filter(n() > 5) %>%
  group_by(cancer_PAM50, gene) %>%
  filter(length(unique(TP53)) == 2) %>%
  ungroup()

#calculate stats
stats <- data %>%
  group_by(cancer_PAM50, gene, TP53) %>%
  summarize(value = list(value)) %>%
  spread(TP53, value) %>%
  mutate(
    median_wt = median(unlist(`TP53 wt`)),
    median_mut = median(unlist(`TP53 mut`)),
    CI95l_wt = t.test(unlist(`TP53 wt`))$conf.int[1],
    CI95l_mut = t.test(unlist(`TP53 mut`))$conf.int[1],
    CI95h_wt = t.test(unlist(`TP53 wt`))$conf.int[2],
    CI95h_mut = t.test(unlist(`TP53 mut`))$conf.int[2],
    pValue = wilcox.test(unlist(`TP53 wt`), unlist(`TP53 mut`), alternative = "two.sided")$p.value
  ) %>%
  ungroup() %>%
  gather(metric, value, -(cancer_PAM50:`TP53 wt`), -pValue) %>%
  separate(col = metric, into = c("metric", "TP53"), sep = "_") %>%
  spread(metric, value) %>%
  mutate(TP53 = if_else(TP53 == "wt", "TP53 wt", "TP53 mut")) %>%
  pFormat(.)

#add fake data where p53 mut is missing (will be hidden in plot) to keep empty space
data <- data %>%
  mutate(TP53 = parse_factor(TP53, levels = c("TP53 wt", "TP53 mut"))) %>%
  mutate(gene = parse_factor(gene, levels = c("lncTAM34a", "miR34a"))) %>%
  mutate(cancer_PAM50 = parse_factor(cancer_PAM50, levels = rev(unique(cancer_PAM50))))

stats <- stats %>%
  mutate(TP53 = parse_factor(TP53, levels = c("TP53 wt", "TP53 mut"))) %>%
  mutate(gene = parse_factor(gene, levels = c("lncTAM34a", "miR34a"))) %>%
  mutate(cancer_PAM50 = parse_factor(cancer_PAM50, levels = rev(unique(cancer_PAM50))))

ggplot(data = NULL) + 
  geom_jitter(
    data = data,
    aes(
      x = cancer_PAM50,
      y = value,
      colour = TP53
    ),
    position = position_dodge(width = 0.75),
    alpha = 0.5
  ) +
  geom_text(
    data = stats, 
    aes(
      x = cancer_PAM50, 
      y = 1, 
      label = pFormat
    )
  ) +
  geom_point(
    data = stats,
    aes(
      x = cancer_PAM50,
      y = median,
      group = TP53
    ),
    colour = "grey30",
    shape = 124,
    size = 4,
    lwd = 1,
    position = position_dodge(width = 0.75)
  ) +
  facet_wrap(~gene, ncol = 2) +
  coord_flip(ylim = c(-12, 2.05)) +
  labs(
    x = "Cancer",
    y = "Expression"
  ) +
  theme_few() +
  scale_colour_economist() +
  theme(legend.position = "top") +
  guides(colour = guide_legend(title = "TP53 status"))

#plot pdf
figure <- ggplot(data = NULL) + 
  geom_jitter(
    data = data,
    aes(
      x = cancer_PAM50,
      y = value,
      colour = TP53
    ),
    position = position_dodge(width = 0.75),
    alpha = 0.5,
    size = 1
  ) +
  geom_text(
    data = stats, 
    aes(
      x = cancer_PAM50, 
      y = 1, 
      label = pFormat
    ),
    size = 3.25
  ) +
  geom_point(
    data = stats,
    aes(
      x = cancer_PAM50,
      y = median,
      group = TP53
    ),
    colour = "grey30",
    shape = 124,
    size = 4,
    lwd = 1,
    position = position_dodge(width = 0.75)
  ) +
  facet_wrap(~gene, ncol = 2) +
  coord_flip(ylim = c(-12, 2.05)) +
  labs(
    x = "Cancer",
    y = "Expression"
  ) +
  theme_few() +
  scale_colour_economist() +
  theme(
    plot.margin = unit(rep(0.1, 4), "lines"),
    legend.position = "top",
    legend.background = element_blank(),
    legend.box.spacing = unit(5, "pt"), #this controls the spacing between strip.text.x and the legend
    legend.margin = margin(rep(0, 4), unit = "pt"),
    legend.key.size = unit(1/5, "cm"),
    legend.title = element_text(size = 9),
    legend.text = element_text(size = 7),
    axis.title.y = element_blank(),
    axis.title.x = element_text(size = 9, margin = margin(t = 5)),
    axis.text.y = element_text(size = 10),
    axis.text.x = element_text(size = 10),
    axis.ticks = element_line(size = 0.25),
    axis.ticks.length = unit(1/15, "cm"),
    panel.border = element_rect(fill = NA, size = 0.15)
  ) +
  guides(colour = guide_legend(title = "TP53 status", override.aes = list(size = 3, alpha = 1)))

##save to pdf
# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 1-Supplement 1b"][[1]])
# ggsave(
#    plot = figure,
#    filename = path,
#    device = cairo_pdf,
#    height = 135,
#    width = 205,
#    units = "mm"
# )

Conclusions

miR34a asRNA and miR34a tend to exhibit a decreased expression in cancer types examined in patients with TP53 mutations.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

We desired to access the polyadenylation status of the miR34a asRNA transcript.

Methods

Cell culture and PCR All cell lines were cultured at 5% CO2 and 37° C with HEK293T cells cultured in DMEM high glucose (Hyclone). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. PCR was then performed and the results analyzed on a 2% agarose gel.

Primers

primers <- data.frame(
    name=c(
        "miR34aAS_F1",
        "miR34aAS_R1",
        "miR34aHG_F",
        "miR34aHG_R",
        "B-actin_F",
        "B-actin_R",
        "U48_F",
        "U48_R"
    ),
    sequence=c(
        "AGC GGC ATC TCC TCC ACC TGA AA",
        "TTG CCT CGT GAG TCC AAG GAG AAT",
        "TCT GCT CCA GTG GCT GAT GAG AAA",
        "GTT CAC TGG CCT CAA AGT TGG CAT",
        "AGG TCA TCA CCA TTG GCA ATG AG",
        "CTT TGC GGA TGT CCA CGT CA",
        "AGT GAT GAT GAC CCC AGG TA",
        "GGT CAG AGC GCT GCG GTG AT"
    )
)
primers

Results

url <- fileMap(type = "png")["Supplementary Figure 2b"][[1]]
knitr::include_graphics(file.path(projectUrl, url))

Conclusions

Results indicate that miR34a asRNA is polyadenylated.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

We investigated the propensity of miR34a asRNA to be alternatively spliced, using PCR cloning and sequencing.

Methods

Cell culture and cloning

All cell lines were cultured at 5% CO2 and 37° C with U2OS cells cultured in McCoy’s 5a (Life Technologies), . All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. PCR was performed with the specified primer sets with annealing temperature 60˚C unless otherwise indicated (see below). PCR was run for 35 cycles for those primer sets with annealing temperature 60˚C and 40 cycles in other cases. Resulting PCR products were run on a agarose gel, gel extracted (8 in total) with the QIAquick gel extraction kit (Qiagen), cloned using the Strata Cloning kit (Agilent), and sequenced. The sequencing results were visualized in the UCSC genome browser. Results aligning to the opposite strand and spurious results (several small 24bp alignments) were removed from the final figure.

Primers

primers <- data.frame(
    name=c(
        "miR34a AS F12",
        "miR34a AS R1",
        "miR34a AS R2",
        "miR34a AS R3",
        "miR34a AS Ex3 R1"
    ),
    sequence=c(
        "AAACACAAGCGTTTACCTGGGTGC",
        "TTGCCTCGTGAGTCCAAGGAGAAT",
        "ATAGGTTCATTTGCCCGATGTGCC",
        "CCACAGCTGTTGCTTCTGAATGCT",
        "TCAGATAGGTACCAAAAATGATGGCCGCAACTAATGACGGAT"
    )
)
primers

Primer combinations:

miR34a AS F12 + miR34a AS R1
miR34a AS F12 + miR34a AS R2
miR34a AS F12 + miR34a AS R3
miR34a AS F12 + miR34a AS Ex3 R1 (56 degrees annealing)
miR34a AS F12 + miR34a AS Ex3 R1 (58 degrees annealing)

Results

url <- fileMap(type = "png")["Supplementary Figure 2c"][[1]]
knitr::include_graphics(file.path(projectUrl, url))

Conclusions

Results indicate that the miR34a asRNA transcript is post-transcriptionally spliced to form several different isoforms.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

asRNAs are known to be capable of exerting regulatory functions on both a transcriptional and post-transcriptional level and therefore, the cellular localization of the asRNA can be informative concerning the method of regulation.

Methods

Quantified RNAseq data from 11 cell lines from the GRCh38 assembly was downloaded from ENCODE and quantifications for miR34a asRNA (ENSG00000234546), ACTB (ENSG00000075624), GAPDH (ENSG00000111640), and MALAT1 (ENSG00000251562) were extracted. Cell lines for which data was downloaded include: A549, GM12878, HeLa-S3, HepG2, HT1080, K562 MCF-7, NCI-H460, SK-MEL-5, SK-N-DZ, SK-N-SH. Initial exploratory analysis revealed that several cell lines should not be included in the final figure for the following reasons: The SK-N-SH has a larger proportion of GAPDH in the nucleus than cytoplasm. The variation of miR34a asRNA expression is too large for SK-MEL-5. K562, HT1080, SK-N-DZ, and NCI-H460 have no or low miR34a asRNA expression. In addition, both the cytoplasmic markers ACTB and GAPDH were analyzed for their ability to differentiate between the nuclear and cytoplasmic fractions, and GAPDH was choosed for the final analysis due its superior performance. Furthermore, only polyadenylated libraries were used in the final analysis, due to the fact that the cellular compartment enrichment was seen to be improved in these samples, and all analyzed genes are reported to be polyadenylated (MALAT1: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2722846/). Only samples with 2 biological replicates were retained. For each cell type, gene, and biological replicate the fraction of transcripts per million (TPM) in each cellular compartment was calculate as the fraction of TPM in the specific compartment by the total TPM. The mean and standard deviation for the fraction was subsequently calculated for each cell type and cellular compartment and this information was represented in the final figure.

Results

data <- getData("Supplementary Figure 2d")

#SK-N-SH has a larger proportion of GAPDH in the nucleus than cytoplasm
#The variation in miR34a asRNA expression is way too large for SK-MEL-5
#K562, HT1080, SK-N-DZ, and NCI-H460 have no or low miR34a asRNA expression
#use GAPDH instead of ACTB for cytoplasmic fraction, seems to be a better representation.

fdata <- data %>% 
  filter(!`Biosample term name` %in% c("K562", "HT1080", "SK-N-DZ", "NCI-H460", "SK-MEL-5", "SK-N-SH")) %>% 
  filter(gene_name != "ACTB") %>%
  select(`Biosample term name`, gene_id, gene_name, fraction, `Library made from`, `Biological replicate(s)`, TPM) %>%
  group_by(`Biosample term name`, gene_id, gene_name, fraction) %>%
  mutate(tmp = n() == 2) %>%
  filter(tmp | `Library made from` == "polyadenylated mRNA") %>%
  ungroup() %>%
  select(-tmp) %>%
  group_by(`Biosample term name`, gene_id, gene_name, `Biological replicate(s)`) %>%
  mutate(frac = TPM / sum(TPM)) %>%
  ungroup()

stats <- fdata %>%
  group_by(`Biosample term name`, gene_id, gene_name, fraction) %>%
  filter(n() == 2) %>%
  summarize(
    n = n(),
    mean = mean(frac),
    sd_min = mean(frac) + sd(frac),
    sd_max = mean(frac) - sd(frac)
  )  %>%
  ungroup()

#prepare for plotting 
stats <- mutate(stats, gene_name = parse_factor(gene_name, levels = c("GAPDH", "MALAT1", "lncTAM34a")))

markdown <- ggplot(data = NULL) +
  geom_point(
    data = stats,
    aes(x = gene_name, y = mean, color = fraction),
    position = position_dodge(width = 1)
  ) +
  facet_wrap(~`Biosample term name`, ncol = 1) +
  geom_errorbar(
    data = stats, 
    aes(x = gene_name, ymax = sd_max, ymin = sd_min, colour = fraction), 
    width = 0,
    position = position_dodge(width = 1)
  ) +
  geom_vline(
    mapping = NULL, 
    xintercept=c(1.5, 2.5), 
    color = "darkgrey", 
    linetype = 3
  ) +
  scale_y_continuous(labels = scales::percent) +
  coord_flip() +
  labs(
    y = "% TPM",
    x = "Gene"
  ) +
  guides(colour = guide_legend(title = "Fraction"))

plotRmarkdown(markdown) +
  theme(
    legend.position = "top",
    axis.title.y = element_text(margin = margin(r = 20, "cm"))
  )

pdf <- ggplot(data = NULL) +
  geom_point(
    data = stats,
    aes(x = gene_name, y = mean, color = fraction),
    position = position_dodge(width = 0.45),
    size = 0.17
  ) +
  facet_wrap(~`Biosample term name`, ncol = 1) +
  geom_errorbar(
    data = stats, 
    aes(x = gene_name, ymax = sd_max, ymin = sd_min, colour = fraction), 
    width = 0,
    position = position_dodge(width = 0.45),
    size = 0.25
  ) +
  geom_vline(
    mapping = NULL,
    xintercept=c(1.5, 2.5), 
    color = "darkgrey", 
    linetype = 3,
    size = 0.35
  ) +
  scale_y_continuous(labels = scales::percent) +
  coord_flip() +
  labs(
    y = "% TPM",
    x = "Gene"
  ) +
  guides(colour = guide_legend(title = "Fraction"))

figure <- plotPDF(pdf) +
  theme(
    axis.text.y = element_text(size = 4),
    panel.spacing = unit(0.1, "lines"),
    strip.text.x = element_text(size = 5, margin = margin(5.5, 5.5, 2.5, 5.5, unit = "pt"))
  )

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 1-Supplement 2e"][[1]])
# ggsave(
#   plot = figure,
#   filename = path,
#   device = cairo_pdf,
#   height = 60,
#   width = 83,
#   units = "mm"
# )

Conclusions

The analysis revealed that the miR34a asRNA transcript localizes to both the nucleus and cytoplasm but primarily resides in the nucleus.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

It is likely, due to the head-to head orientation of miR34a HG and asRNA, that transcription initiation may be activated from a single promoter in a bi-directional manner. To investigate whether miR34a HG and asRNA are transcribed from the same promoter as convergent transcripts, we cloned the miR34a HG promoter, including the p53 binding site, into a luciferase/renilla dual reporter vector which we hereafter refer to as p1. The p1 sequence was previously published in Raver-Shapira, N., et al., Transcriptional activation of miR-34a contributes to p53-mediated apoptosis. Mol Cell, 2007. 26(5): p. 731-43.

Methods

Cell culture, transfection and luminescence quantification

All cell lines were cultured at 5% CO2 and 37° C with HEK293T cells cultured in DMEM high glucose (Hyclone) and HCT116 cells in McCoy’s 5a (Life Technologies). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. Cells were plated at 10,000 cells per well in a 96-well plate with a white bottom and cultured overnight. The following day cells were co-transfected with 10ng of empty, p1, or p2 plasmid and GFP using the standard lipofektamine 2000 (Life Technologies) protocol. The expression of GFP and luminescence was measured 24 h post transfection by using the Dual-Glo Luciferase Assay System (Promega) and detected by the GloMax-Multi+ Detection System (Promega). The expression of luminescence was normalized to GFP.

Analysis

Three independant experiments (biological replicates) were performed in total. The Student’s t-test was used to compare the empty vs p1 group’s log2 activity for both genes.

Constructs

ps <- data.frame(
  name = c("p1"),
  sequence = c("GCG CCC TGC CTG GCC CCC ACC TGG TCC TCT TTC CTT TTC AGG TGG AGG AGA TGC CGC TGT
              CCC GTC GGT CTG GGG ACA GCC CAG CTC CCC GGA TCC CGG GCT GGA GAG ACG CGT CGC GGC   
              CCC GGG GCC TGG TGG CAC GAG CAG GAA GGA GGA CCC GGC GGC GGG CTC TGC CTG GGC TTG 
              CCT GGG CTT GTT CCG AGC CGG GCT GCT TCT CGG TGA CCA CGC AGA TCG GGG GCA TTT GGA
              GAT TTT GCG GGA GTC CTG CAG CCA AGC TCC GGG GCA GGA GAG GCC TGG AAG CCT GCA CTA
              CCT GCT C"
  )
)
ps

Primers

primers <- data.frame(
    name=c(
        "Luc_set_II_F",
        "Luc_set_II_R",
        "Renilla_pBiDir_F1",
        "Renilla_pBiDir_R1",
        "B-actin_F",
        "B-actin_R"
    ),
    sequence=c(
        "AAG ATT CAA AGT GCG CTG CTG",
        "TTG CCT GAT ACC TGG CAG ATG",
        "TAA CGC GGC CTC TTC TTA TTT",
        "GAT TTG CCT GAT TTG CCC ATA",
        "AGG TCA TCA CCA TTG GCA ATG AG",
        "CTT TGC GGA TGT CCA CGT CA"
    )
)
primers

Results

url <- fileMap(type = "png")["Supplementary Figure 3a"][[1]]
knitr::include_graphics(file.path(projectUrl, url))

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

It is likely, due to the head-to head orientation of miR34a HG and asRNA, that transcription initiation may be activated from a single promoter in a bi-directional manner. To investigate whether miR34a HG and asRNA are transcribed from the same promoter as convergent transcripts, we cloned the miR34a HG promoter, including the p53 binding site, into a luciferase/renilla dual reporter vector which we hereafter refer to as p1. The p1 sequence was previously published in Raver-Shapira, N., et al., Transcriptional activation of miR-34a contributes to p53-mediated apoptosis. Mol Cell, 2007. 26(5): p. 731-43.

Methods

Cell culture, transfection and luminescence quantification

All cell lines were cultured at 5% CO2 and 37° C with HEK293T cells cultured in DMEM high glucose (Hyclone) and HCT116 cells in McCoy’s 5a (Life Technologies). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. Cells were plated at 10,000 cells per well in a 96-well plate with a white bottom and cultured overnight. The following day cells were co-transfected with 10ng of empty, p1, or p2 plasmid and GFP using the standard lipofektamine 2000 (Life Technologies) protocol. The expression of GFP and luminescence was measured 24 h post transfection by using the Dual-Glo Luciferase Assay System (Promega) and detected by the GloMax-Multi+ Detection System (Promega). The expression of luminescence was normalized to GFP.

Analysis

Three independant experiments (biological replicates) were performed in total. The Student’s t-test was used to compare the empty vs p1 group’s log2 activity for both genes.

Constructs

ps <- data.frame(
  name = c("p1"),
  sequence = c("GCG CCC TGC CTG GCC CCC ACC TGG TCC TCT TTC CTT TTC AGG TGG AGG AGA TGC CGC TGT
              CCC GTC GGT CTG GGG ACA GCC CAG CTC CCC GGA TCC CGG GCT GGA GAG ACG CGT CGC GGC   
              CCC GGG GCC TGG TGG CAC GAG CAG GAA GGA GGA CCC GGC GGC GGG CTC TGC CTG GGC TTG 
              CCT GGG CTT GTT CCG AGC CGG GCT GCT TCT CGG TGA CCA CGC AGA TCG GGG GCA TTT GGA
              GAT TTT GCG GGA GTC CTG CAG CCA AGC TCC GGG GCA GGA GAG GCC TGG AAG CCT GCA CTA
              CCT GCT C"
  )
)
ps

Primers

primers <- data.frame(
    name=c(
        "Luc_set_II_F",
        "Luc_set_II_R",
        "Renilla_pBiDir_F1",
        "Renilla_pBiDir_R1",
        "B-actin_F",
        "B-actin_R"
    ),
    sequence=c(
        "AAG ATT CAA AGT GCG CTG CTG",
        "TTG CCT GAT ACC TGG CAG ATG",
        "TAA CGC GGC CTC TTC TTA TTT",
        "GAT TTG CCT GAT TTG CCC ATA",
        "AGG TCA TCA CCA TTG GCA ATG AG",
        "CTT TGC GGA TGT CCA CGT CA"
    )
)
primers

Results

url <- fileMap(type = "png")["Supplementary Figure 3a"][[1]]
knitr::include_graphics(file.path(projectUrl, url))

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

Functional characterization of individual antisense transcripts has previously shown their capability to regulate their sense gene. We therefore investigated the possibility that miR34a asRNA may regulate miR34a HG levels. We hypothesized that the overlapping regions of the sense and antisense transcripts may have a crucial role in miR34a asRNAs ability to regulate miR34a HG via RNA:DNA or RNA:RNA interaction. Accordingly, we first co-transfected the p1 construct, containing the overlapping region of the two transcripts, and an short hairpin (sh) RNA targeting renilla into HEK293T cells .

Methods

We utilized the P1 construct where the overlapping region of miR34a HG and miR34a AS is cloned with luciferase downstream of miR34a HG and renilla downstream of miR34a AS. The p1 sequence was previously published in Raver-Shapira, N., et al., Transcriptional activation of miR-34a contributes to p53-mediated apoptosis. Mol Cell, 2007. 26(5): p. 731-43. All cell lines were cultured at 5% CO2 and 37° C with HEK293T cells cultured in DMEM high glucose (Hyclone). All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. 2.5x10^5 293T cells were seeded in a 12-well plate. After 24hrs these were co-transfected with 50 ng of the P1 construct and 250 ng shRenilla using lipo2k standard protocol. After 48 hours, RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. QPCR was carried out using KAPA 2G SYBRGreen (Kapa Biosystems) using the Applied Biosystems 7900HT machine with the cycling conditions: 95 °C for 3 min, 95 °C for 3 s, 60 °C for 30 s.

Constructs

p1 <- data.frame(
    name="p1",
    sequence="GCG CCC TGC CTG GCC CCC ACC TGG TCC TCT TTC CTT TTC AGG TGG AGG AGA TGC CGC TGT
              CCC GTC GGT CTG GGG ACA GCC CAG CTC CCC GGA TCC CGG GCT GGA GAG ACG CGT CGC GGC   
              CCC GGG GCC TGG TGG CAC GAG CAG GAA GGA GGA CCC GGC GGC GGG CTC TGC CTG GGC TTG 
              CCT GGG CTT GTT CCG AGC CGG GCT GCT TCT CGG TGA CCA CGC AGA TCG GGG GCA TTT GGA
              GAT TTT GCG GGA GTC CTG CAG CCA AGC TCC GGG GCA GGA GAG GCC TGG AAG CCT GCA CTA
              CCT GCT C"
)
p1

shRNAs

shRNAs <- data.frame(
    name=c(
        "shRenilla1.1",
        "shRenilla2.1"
    ),
    sequence=c(
        "AAT ACA CCG CGC TAC TGG C",
        "TAA CGG GAT TTC ACG AGG C"
    )
)
shRNAs

Primers

primers <- data.frame(
    name=c(
        "Luc_set_II_F",
        "Luc_set_II_R",
        "Renilla_pBiDir_F1",
        "Renilla_pBiDir_R1",
        "B-actin_F",
        "B-actin_R"
    ),
    sequence=c(
        "AAG ATT CAA AGT GCG CTG CTG",
        "TTG CCT GAT ACC TGG CAG ATG",
        "TAA CGC GGC CTC TTC TTA TTT",
        "GAT TTG CCT GAT TTG CCC ATA",
        "AGG TCA TCA CCA TTG GCA ATG AG",
        "CTT TGC GGA TGT CCA CGT CA"
    )
)
primers

Results

data <- getData("Supplementary Figure 3c")

#Luciferase = miR34a HG
#Renilla = miR34a AS

#calculate dct
groups <- c("shRNA", "Biological Replicate", "belongsTo", "gene")
gois <- c("Luciferase", "Renilla")

data <- data %>%
  technicalMeans(., groups) %>%
  dct(., gois, Actin, groups, F)

#calculate ddct
groups[groups == "gene"] <- "GOI"
logical <- tibble(
  `shRNA` = pull(data, shRNA) == "shCtrl"
)

data <- data %>%
  ddct(., logical, groups) %>%
  folds(.)

#calculate stats
stats <- data %>%
  group_by(shRNA, belongsTo, GOI) %>%
  summarize(
    mean = mean(log2fold),
    CI95l = t.test(log2fold)$conf.int[1],
    CI95h =  t.test(log2fold)$conf.int[2]
  )

#plot for rmarkdown

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = shRNA,
      y = log2fold,
      fill = GOI,
      group = interaction(shRNA, GOI)
    ),
    binwidth = 1/10,
    colour = NA,
    alpha = 0.75,
    position = position_dodge(width = 0.9), 
    binaxis = 'y', 
    stackdir = 'center',
    dotsize = 3,
    show.legend = TRUE
  ) +
  facet_grid(. ~ belongsTo, scales = "free") +
  scale_y_continuous("\u0394\u0394Ct")+
  labs(x = "shRNA")

plotRmarkdown(p) +
  geom_linerange(
    data = stats,
    aes(
      x = shRNA, 
      ymin = CI95l, 
      ymax = CI95h, 
      group = GOI
    ),
    colour = "gray20",
    position = position_dodge(width = 0.9),
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = shRNA, 
      y = mean, 
      group = GOI
    ),
    colour = "gray20",
    position = position_dodge(width = 0.9),
    shape = 95,
    size = 6,
    show.legend = FALSE
  ) +
  theme(
    strip.text.x = element_blank(),
    axis.text.x = element_text(angle = 90)
  )

figure <- plotPDF(p) +
  geom_linerange(
    data = stats,
    aes(
      x = shRNA, 
      ymin = CI95l, 
      ymax = CI95h, 
      group = GOI
    ),
    colour = "grey30",
    position = position_dodge(width = 0.9),
    size = 0.3,
    show.legend = FALSE
  ) +
  geom_point(
    data = stats,
    aes(
      x = shRNA, 
      y = mean, 
      group = GOI
    ),
    colour = "grey30",
    position = position_dodge(width = 0.9),
    shape = 95,
    size = 4,
    show.legend = FALSE
  ) +
  theme(
    strip.text.x = element_blank(),
    axis.text.x = element_text(angle = 90)
  )

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 2-Supplement 2"][[1]])
# ggsave(
#    plot = figure,
#    filename = path,
#    device = cairo_pdf,
#    height = 60,
#    width = 83,
#    units = "mm"
# )

Conclusions

Analysis of luciferase and renilla expression revealed that by decreasing levels of the renilla transcript (corresponding to miR34a asRNA) luciferase (corresponding to miR34a HG) levels were concomitantly decreased. These results indicate that miR34a asRNA is capable of regulating the miR34a HG in a concordant manner and that the overlapping region of the miR34a asRNA may be sufficient to mediate this regulation.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

miR34a asRNA over-expressing cell lines, previously generated, showed a high overexpression level of miR34a asRNA despite a low multiplicity of infection. This is most likley due to the fact that all infected cell lines have practically no endogenous miR34a asRNA expression. To investigate the physiological relevance of the over-expression the miR34a asRNA levels were compared to the expression levels in HEK293t cells, which have a high endogenous miR34a expression level.

Methods

miR34a asRNA cloning

RNA was harvested from U2OS cells using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. PCR was run using the miR34aAS_cloning_F4 and miR34aAS_cloning_Ex3_1 primers for 40 cycles at annealing temperature 58˚C and run on an agarose gel. Bands were excised and extracted using the QIAquick gel extraction kit (Qiagen). PCR products were cloned using the Strata Clone kit (Aligent) and sequenced (results below). The insert for the lentiviral construct (p-HIV-I2AG-U6) was subsequently sub-cloned from the sequenced plasmid using primers with PacI and NheI cleavage sites. The cloned transcript largley mirrors the GENCODE annotated transcript.

Cloning primers

primers <- data.frame(
    name=c(
        "miR34aAS_cloning_F4",
        "miR34aAS_cloning_Ex3_1"
    ),
    sequence=c(
        "ACGCGTCTCTCCAGCCCGGGAT",
        "AATGATGGCCGCAACTAATGACGG"
    )
)
primers

Sequencing results

construct <- data.frame(
    name="miR34a_asRNA",
    sequence="ACG CGT CTC TCC AGC CCG GGA TCC GGG GAG CTG GGC TGT CCC CAG ACC GAC GGG ACA GCG GCA
                TCT CCT CCA CCT GAA AAG GAA AGA GGA CCA GTT TGC AGG ACT CCG AAC TGG GCC CGC GAG ATC 
                TCC ACC TGC GCA AAA CGA AAG GGC GGA TTC TCC TTG GAC TCA CGA GGC AAC CGC TCC CCG GGG 
                TGA GAA CGG GGG ACT CAT TCC TCC GGC ACT GGG AGA AGA CGA TTC TTC AGG AGG AGG ACA GGG 
                AAG CGA ATG CTA CCC AGA TGT CTC AGT ATA CTG GCT CGC GGC ACA TCG GGC AAA TGA ACC TAT 
                CAG ATA ACA ACG GCA GAT CAG ATG CCT GAG CAT TCA GAA GCA ACA GCT GTG GAG CCC CCG TGG 
                GTT CAG AAG GCC TGG TTC CCG TCT CCA GAA GCC TGG CTC TCC TCC CTC CTG GGC CCA CTA CTT 
                TGG CTT CTT GTT CCT ACG TAC AAG GAG TTG CGA AGA AGG CAA CTC TTC CCC TCC CTG AAG CCA
                AAG GAA TGA AAC AGA CTA GGG CGG GAG AGG TGG CCA TCC GTC ATT AGT TGC GGC CAT CA"
)
construct

Lentivirus production and infection

HEK293T cells were transfected with viral and expression constructs using Lipofectamine 2000 (Life Technologies), after which viral supernatants were harvested 48 and 72 hours post-transfection. Viral particles were concentrated using PEG-IT solution (Systems Biosciences) according to the manufacturer’s recommendations. HEK293T cells were used for virus titration and GFP expression was evaluated 72hrs post-infection via flow cytometry after which TU/ml was calculated. Cell lines were subsequently infected with the lentiviral particals using a multuplicity of infection = 1. After 48 hours, cells were placed under mycophenolic acid selection and the selection process was monitored using GFP expression via flow cytometry. When GFP expression was present in >90% of the cells, cells were frozen for downstream analysis.

QPCR

RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. QPCR was carried out using KAPA 2G SYBRGreen (Kapa Biosystems) using the Applied Biosystems 7900HT machine with the cycling conditions: 95 °C for 3 min, 95 °C for 3 s, 60 °C for 30 s. QPCR for miRNA expression analysis was performed according to the protocol for the TaqMan microRNA Assay kit (Life Technologies) with primer/probe sets purchased from the same company (TaqMan® MicroRNA Assay, hsa-miR-34a, human and Control miRNA Assay, RNU48, human) and the same cycling scheme as above.

Analysis

Two experimental (technical) replicates were included in each QPCR run and delta ct was calculated for each sample using the mean of the gene of interest’s technical replicates and the house keeping gene’s technical replicates. Delta-delta ct was calculated for each sample by subtracting the median dct value for the corresponding untreated samples. Five independant experiments were performed in total for all cell lines.

Primers

primers <- data.frame(
    name=c(
        "ß-actin Fwd",
        "ß-actin Rev",
        "miR34a_asF1",
        "miR34a_asR1"
        
    ),
    sequence=c(
        "AGGTCATCACCATTGGCAATGAG",
        "CTTTCGGGATGTCCACGTCA",
        "AGCGGCATCTCCTCCACCTGAAA",
        "TTGCCTCGTGAGTCCAAGGAGAAT"
        
    )
)
primers

Results

data <- getData("Supplementary Figure 4a")

#remove PC3 experiment 6, B-act mean Ct > 30, too high.
data <- filter(data, `Biological Replicate` != 6 | `Cell line` != "PC3")

#calculate dct
groups <- c("Condition", "Biological Replicate", "Cell line", "gene")
gois <- c("lncTAM34a")

data <- technicalMeans(data, groups) %>%
  dct(., gois, Actin, groups)

#calculate ddct
groups[groups == "gene"] <- "GOI"
logical <- tibble(
  `Cell line` = data$`Cell line` == "HEK293t",
  Condition = data$Condition == "wt"
)

data <- data %>%
  ddct(., logical, groups) %>%
  folds(.)

#calculate stats
stats <- data %>%
  group_by(`Cell line`, Condition) %>%
  summarize(
    n = n(),
    mean = mean(log2fold),
    CI95l = t.test(log2fold)$conf.int[1],
    CI95h = t.test(log2fold)$conf.int[2]
  ) %>%
  ungroup()

p <- ggplot(data = NULL)+
  geom_dotplot(
    data = data, 
    aes(
      x = Condition, 
      y = log2fold, 
      fill = Condition
    ),
    colour = NA,
    binwidth = 1/10,
    alpha = 0.75,
    position = position_dodge(width = 0.9), 
    binaxis='y', 
    stackdir='center',
    dotsize = 6,
    show.legend = TRUE
  ) +
  facet_grid(. ~ `Cell line`, scales = "free") +
  labs(
    x = "Cell line",
    y = "log2(Fold)",
    title = "miR34a asRNA"
  ) +
  guides(fill = FALSE)

markdown <- p +
  geom_point(
    data = stats,
    aes(
      x = Condition, 
      y = mean
    ),
    colour = "grey20",
    show.legend = FALSE,
    shape = 95,
    size = 6
  ) +
  geom_linerange(
    data = stats,
    aes(
      x = Condition, 
      ymin = CI95l, 
      ymax = CI95h
    ),
    colour = "gray20",
    show.legend = FALSE
  )

plotRmarkdown(markdown)

pdf <- p +
  geom_point(
    data = stats,
    aes(
      x = Condition, 
      y = mean
    ),
    colour = "gray30",
    show.legend = FALSE,
    size = 4,
    shape = 95
  ) +
  geom_linerange(
    data = stats,
    aes(
      x = Condition, 
      ymin = CI95l, 
      ymax = CI95h
    ),
    colour = "gray30",
    size = 0.3,
    show.legend = FALSE
  )

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 3-Supplement 1"][[1]])
# ggsave(
#   plot = figure,
#    filename = path,
#   device = cairo_pdf,
#   height = 60,
#   width = 160,
#   units = "mm"
# )

Conclusions

Results indicate that miR34a asRNA is over-expressed 13-22 fold greater that endogenous HEK293t levels.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

To further investigate the characteristics of the miR34a asRNA stable lines, we measured the expression level of known miR34a target cyclin D1. miR34a has been previously shown to inhibit cell cycle progression, partially via its regulation of cyclin D1.

Methods

Cell culture and QPCR

All cell lines were cultured at 5% CO2 and 37° C with PC3 cells in RPMI (Hyclone) and 2 mM L-glutamine. All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. Cells were seeded at 2x10^5 cells per well in a 12-well plate. After 24 hours, RNA was extracted using the RNeasy mini kit (Qiagen) and subsequently treated with DNase (Ambion Turbo DNA-free, Life Technologies). 500ng RNA was used for cDNA synthesis using MuMLV (Life Technologies) and a 1:1 mix of oligo(dT) and random nanomers. QPCR was carried out using KAPA 2G SYBRGreen (Kapa Biosystems) using the Applied Biosystems 7900HT machine with the cycling conditions: 95 °C for 3 min, 95 °C for 3 s, 60 °C for 30 s.

Analysis

Two technical replicates were included for each sample and each QPCR run. Delta ct was calculated by subtracting the mean of the technical replicates for each sample from the mean value for the technical replicates from the house keeping gene for the corresponding sample. Delta-delta ct was calculated for each sample by subtracting the median dct value for the corresponding mock samples. Six independant experiments (biological replicates) were performed in total. The Student’s t-test was used to compare the mock and miR34a asRNA overexpressing group’s delta-delta ct values with the null hypothesis that there was no difference in the two group’s means.

Primers

primers <- data.frame(
    name=c(
        "CCND1 Fwd",    
        "CCND1 Rev",
        "B-actin_F",
        "B-actin_R"
    ),
    sequence=c(
        "CGT GGC CTC TAA GAT GAA GG",
        "CTG GCA TTT TGG AGA GGA AG",
        "AGG TCA TCA CCA TTG GCA ATG AG",
        "CTT TGC GGA TGT CCA CGT CA"
    )
)
primers

Results

data <- getData("Supplementary Figure 4b")

#calculate dct
groups <- c("condition", "Biological Replicate", "Cell line", "gene")
gois <- c("CCND1")

data <- data %>%
  technicalMeans(., groups) %>%
  dct(., gois, Actin, groups, F)

#calculate ddct
groups[groups == "gene"] <- "GOI"
logical <- tibble(
  condition = pull(data, condition) == "mock"
)
 
data <- data %>%
  ddct(., logical, groups) %>%
  folds(.)

#calculate stats
stats <- calcStats(data, condition, "mock", groups) %>%
  pFormat(.)

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = condition,
      y = log2fold,
      fill = condition
    ),
    binwidth = 1/200,
    colour = NA,
    #alpha = 0.5,
    position = position_dodge(width = 0.9), 
    binaxis = 'y', 
    stackdir = 'center',
    dotsize = 4,
    show.legend = TRUE
  ) +
  scale_y_continuous("\u0394\u0394Ct") +
  labs(
    x = "Cell line",
    title = "CCND1"
  ) +
  guides(fill = FALSE)

markdown <- p + 
  geom_linerange(
    data = stats,
    aes(
      x = condition,
      ymin = CI95l,
      ymax = CI95h
    ),
    colour = "gray20"
  ) +
  geom_point(
    data = stats,
    aes(
      x = condition,
      y = mean
    ),
    colour = "gray20",
    size = 6,
    shape = 95
  ) +
  geom_segment(
    aes(
      x = 1,
      y = 0.175, 
      xend = 2,
      yend = 0.175
    ),
    colour = "grey43"
  ) +
  geom_label(
    data = stats,
    aes(
      x = 1.5,
      y = 0.19,
      label = pFormat
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 5,
    family = "Arial Unicode MS"
  )
  
plotRmarkdown(markdown)

pdf <- p +
  geom_linerange(
    data = stats,
    aes(
      x = condition,
      ymin = CI95l,
      ymax = CI95h
    ),
    colour = "grey30",
    size = 0.3
  ) +
  geom_point(
    data = stats,
    aes(
      x = condition,
      y = mean
    ),
    colour = "grey30",
    shape = 95,
    size = 4
  ) +
  geom_segment(
    aes(
      x = 1,
      y = 0.175, 
      xend = 2,
      yend = 0.175
    ),
    colour = "grey30",
    size = 0.3
  ) +
  geom_label(
    data = stats,
    aes(
      x = 1.5,
      y = 0.2,
      label = pFormat
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    colour = "grey30",
    size = 2.5,
    family = "Arial Unicode MS"
  )

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 3-Supplement 2a"][[1]])
# ggsave(
#   plot = figure,
#   filename = path,
#   device = cairo_pdf,
#   height = 60,
#   width = 43,
#   units = "mm"
# )

Conclusions

CCND1 expression levels are decreased in PC3 cell lines stably over-expressing miR34a asRNA.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1

Introduction

After observing decreased mRNA levels for the miR34a target CCND1, we were interested whether this has also consequences for its protein levels.

Methods

Cell culture and western blot All cell lines were cultured at 5% CO2 and 37° C with PC3 cells in RPMI (Hyclone) and 2 mM L-glutamine. All growth mediums were supplemented with 10% heat-inactivated FBS and 50 μg/ml of streptomycin and 50 μg/ml of penicillin. We seeded three independent experiments of 150,000 cells/well in a six well plate (PC3 mock vs. PC3 miR34a AS F4) and grew them for 24 hours. Before harvesting cells were controlled for their confluency. Cells with a confluence of 60-75%, were harvested using trypsin. Cell pellets were frozen down at -80°C before lysed for Western Blot analysis. Cells were lysed in RIPA Buffer and run on a 4-12% Bis-Tris-SDS Gel using MOPS running buffer. Proteins were transferred onto a nitrocellulose membrane using iBlot Turbo Blotting Device. Proteins were transferred for 7 minutes. Membranes were blocked for 1 hour at room temperature in 5% milk. Membranes were incubated with a cyclin D1 antibody (92G2 Rabbit mAb, Cell Signalling, 1:1000) overnight at 4°C. The membrane was incubated using a goat anti-rabbit antibody conjugated to HRP for 1 hour at room temperature. Membranes were washed 3-times for 5min each in TBS-T. Membranes were developed using chemiluminescence. After the picture was taken, the membrane was stripped using 0.4M NaOH solution for 20min. Blocking step was repeated and primary antibody against GAPDH (Abcam #ab9485, 1:5000) was incubated for overnight at 4°C. Consequently, the membrane was incubated using a goat anti-rabbit antibody conjugated to HRP for 1 hour at room temperature. Membranes were washed 3-times for 5min in TBS-T. Membranes were developed using chemiluminescence.

Analysis

Resulting bands on the western blots were quantified using the ImageJ software with three parameters being measured, intensity, background, and pixel. The quantative value was derived for each band by first subtracting the background from the intensity and subsequently multiplying this value by the pixel value. The fraction of CCND1 over the house keeping gene was then calculated using each sample’s values and the corresponding house keeping genes values. Fold change was then calculated by dividing the normalized CCND1 values in the miR34a asRNA overexpressing samples by the median of the corresponding mock samples. Fold change values were subsequently log2 transformed and the Student’s t-test was utilized to compare mock vs. miR34a asRNA overexpressing samples using three independant experiments with the null hypothesis that the mean values were the same in each group.

Results

Western blot

url <- "stable_line_ccnd1_prot/stable_line_ccnd1_prot.png"
knitr::include_graphics(file.path(projectUrl, url))

Quantification

data <- getData("Supplementary Figure 4c")

#normalize to house keeping gene
data <- data %>%
  mutate(value = (Intensity - background) * pixel) %>%
  select(-pixel, -Intensity, -background) %>%
  spread(protein, value) %>%
  mutate(normHK = CCND1 / GAPDH)

#normalize to mock
cnt <- filter(data, condition == "mock") %>% 
  pull(normHK) %>% 
  median

data <- mutate(data, log2fold = log2(normHK / cnt))

#calculate stats
stats <- data %>%
  group_by(condition) %>%
  summarize(
    n = n(),
    mean = mean(log2fold),
    CI95l = t.test(log2fold)$conf.int[1],
    CI95h = t.test(log2fold)$conf.int[2],
    pValue = t.test(log2fold, pull(filter(data, condition == "mock"), log2fold))$p.value
  ) %>%
  pFormat(.)

p <- ggplot(data = NULL) +
  geom_dotplot(
    data = data, 
    aes(
      x = condition,
      y = log2fold,
      fill = condition
    ),
    binwidth = 1/30,
    colour = NA,
    #alpha = 0.5,
    position = position_dodge(width = 0.9), 
    binaxis = 'y', 
    stackdir = 'center',
    dotsize = 4,
    show.legend = TRUE
  ) +
  scale_y_continuous("log2(Fold)") +
  labs(
    x = "Cell line",
    title = "CCND1 protein"
  ) +
  guides(fill = FALSE)

markdown <- p +
  geom_point(
    data = stats,
    aes(
      x = condition,
      y = mean
    ),
    colour="gray20",
    size = 6,
    shape = 95
  ) +
  geom_segment(
    data = stats,
    aes(
      x = 1,
      y = max(CI95h) + 0.25,
      xend = 2,
      yend = max(CI95h) + 0.25
    ),
    colour="grey43"
  ) +
  geom_label(
    data = stats,
    aes(
      x = 1.5,
      y = max(CI95h) + 0.4,
      label = pFormat
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    size = 5,
    family = "Arial Unicode MS"
  ) +
   geom_linerange(
    data = stats,
    aes(
      x = condition,
      ymin = CI95l,
      ymax = CI95h
    ),
    colour="gray20"
  )

plotRmarkdown(markdown)

pdf <- p + 
  geom_point(
    data = stats,
    aes(
      x = condition,
      y = mean
    ),
    colour="gray30",
    shape = 95,
    size = 4
  ) +
  geom_segment(
    data = stats,
    aes(
      x = 1,
      y = max(CI95h) + 0.25,
      xend = 2,
      yend = max(CI95h) + 0.25
    ),
    colour="gray30",
    size = 0.3
  ) +
  geom_label(
    data = stats,
    aes(
      x = 1.5,
      y = max(CI95h) + 0.4,
      label = pFormat
    ),
    label.size = 0,
    label.padding = unit(0.01, "lines"),
    show.legend = FALSE,
    fill = "white",
    colour = "gray30",
    size = 2.5,
    family = "Arial Unicode MS"
  ) +
   geom_linerange(
    data = stats,
    aes(
      x = condition,
      ymin = CI95l,
      ymax = CI95h
    ),
    colour="gray30",
    size = 0.3
  )

figure <- plotPDF(pdf)

# path <- file.path("~/GitHub/miR34a_asRNA_project/inst", fileMap(type = "pdf")["Figure 3-Supplement 2b"][[1]])
# ggsave(
#   plot = figure,
#   filename = path,
#   device = cairo_pdf,
#   height = 60,
#   width = 43,
#   units = "mm"
# )

Conclusions

Cyclin D1 protein levels are significantly reduced in PC3 cells overexpressing miR34a asRNA, consistent with increased miR34a expression levels and the reduced G1 cell cycle phase observed in PC3 cells.

sessionInfo()

## R version 3.4.3 (2017-11-30)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS High Sierra 10.13.3
## 
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] grid      stats     graphics  grDevices utils     datasets  methods  
## [8] base     
## 
## other attached packages:
##  [1] KMsurv_0.1-5                  survival_2.41-3              
##  [3] broom_0.4.3                   bindrcpp_0.2                 
##  [5] gtable_0.2.0                  ggthemes_3.4.0               
##  [7] printr_0.1                    forcats_0.2.0                
##  [9] stringr_1.2.0                 dplyr_0.7.4                  
## [11] purrr_0.2.4                   readr_1.1.1                  
## [13] tidyr_0.8.0                   tibble_1.4.2                 
## [15] ggplot2_2.2.1.9000            tidyverse_1.2.1              
## [17] knitr_1.19                    miR34AasRNAproject_0.0.0.0001
## 
## loaded via a namespace (and not attached):
##  [1] Rcpp_0.12.15           lubridate_1.7.1        lattice_0.20-35       
##  [4] prettyunits_1.0.2      assertthat_0.2.0       rprojroot_1.3-2       
##  [7] digest_0.6.15          psych_1.7.8            R6_2.2.2              
## [10] cellranger_1.1.0       plyr_1.8.4             backports_1.1.2       
## [13] evaluate_0.10.1        highr_0.6              httr_1.3.1            
## [16] pillar_1.1.0.9000      rlang_0.2.0.9000       progress_1.1.2.9002   
## [19] lazyeval_0.2.1         curl_3.1               readxl_1.0.0          
## [22] rstudioapi_0.7         gggenes_0.2.0.9003     Matrix_1.2-12         
## [25] rmarkdown_1.9          liftr_0.8              labeling_0.3          
## [28] splines_3.4.3          selectr_0.3-1          foreign_0.8-69        
## [31] ansistrings_1.0.0.9000 munsell_0.4.3          rgeolocate_1.0.1      
## [34] compiler_3.4.3         modelr_0.1.1           pkgconfig_2.0.1       
## [37] mnormt_1.5-5           htmltools_0.3.6        tidyselect_0.2.3      
## [40] ggfittext_0.5.0        XML_3.98-1.9           crayon_1.3.4          
## [43] withr_2.1.1.9000       nlme_3.1-131           jsonlite_1.5          
## [46] magrittr_1.5           scales_0.5.0.9000      cli_1.0.0.9001        
## [49] stringi_1.1.6          reshape2_1.4.3         xml2_1.2.0            
## [52] RColorBrewer_1.1-2     rematch2_2.0.1         tools_3.4.3           
## [55] glue_1.2.0             hms_0.4.1              parallel_3.4.3        
## [58] yaml_2.1.16            colorspace_1.3-2       rvest_0.3.2           
## [61] bindr_0.1              haven_1.1.1